Importance: Thoracentesis is performed to identify the cause of a pleural effusion. Although generally safe, thoracentesis may be complicated by transient hypoxemia, bleeding, patient discomfort, reexpansion pulmonary edema, and pneumothorax.
Objective: To identify the best means for differentiating between transudative and exudative effusions and also to identify thoracentesis techniques for minimizing the risk of complications by performing a systematic review the evidence.
Data sources: We searched The Cochrane Library, MEDLINE, and Embase from inception to February 2014 to identify relevant studies.
Study selection: We included randomized and observational studies of adult patients undergoing thoracentesis that examined diagnostic tests for differentiating exudates from transudates and evaluated thoracentesis techniques associated with a successful procedure with minimal complications.
Data extraction and synthesis: Two investigators independently appraised study quality and extracted data from studies of laboratory diagnosis of pleural effusion for calculation of likelihood ratios (LRs; n = 48 studies) and factors affecting adverse event rates (n = 37 studies).
Results: The diagnosis of an exudate was most accurate if cholesterol in the pleural fluid was greater than 55 mg/dL (LR range, 7.1-250), lactate dehydrogenase (LDH) was greater than 200 U/L (LR, 18; 95% CI, 6.8-46), or the ratio of pleural fluid cholesterol to serum cholesterol was greater than 0.3 (LR, 14; 95% CI, 5.5-38). A diagnosis of exudate was less likely when all Light's criteria (a ratio of pleural fluid protein to serum protein >0.5, a ratio of pleural fluid LDH to serum LDH >0.6, or pleural fluid LDH >two-thirds the upper limit of normal for serum LDH) were absent (LR, 0.04; 95% CI, 0.02-0.11). The most common complication of thoracentesis was pneumothorax, which occurred in 6.0% of cases (95% CI, 4.0%-7.0%). Chest tube placement was required in 2.0% of procedures (95% CI, 0.99%-2.9%) in which a patient was determined to have radiographic evidence of a pneumothorax. With ultrasound, a radiologist's marking the needle insertion site was not associated with decreased pneumothorax events (skin marking vs no skin marking odds ratio [OR], 0.37; 95% CI, 0.08-1.7). Use of ultrasound by any experienced practitioner also was not associated with decreased pneumothorax events (OR, 0.55; 95% CI, 0.06-5.3).
Conclusions and relevance: Light's criteria, cholesterol and pleural fluid LDH levels, and the pleural fluid cholesterol-to-serum ratio are the most accurate diagnostic indicators for pleural exudates. Ultrasound skin marking by a radiologist or ultrasound-guided thoracentesis were not associated with a decrease in pneumothorax events.