| |
著者: José Manuel Porcel, Manuel Vives, Aureli Esquerda, Agustín Ruiz
雑誌名: Respir Med. 2006 May;100(5):933-7. doi: 10.1016/j.rmed.2005.06.017. Epub 2005 Oct 25.
Abstract/Text
AIM: To assess the value of the British Thoracic Society (BTS) and the American College of Chest Physicians (ACCP) guidelines to predict which patients with non-purulent parapneumonic effusions (PPE) warrant chest tube drainage.
METHODS: A retrospective chart review was performed on all patients who underwent thoracentesis because of a PPE over a 10-year period at a Spanish medical center. Classification of PPE as complicated (CPPE) or uncomplicated (UPPE) was based on the clinician's decision to insert a chest tube to resolve the effusion. Empyema was defined as pus in the pleural space. Data collected included patient demographics, size of the effusion, and microbiological and pleural fluid chemistries that might influence the physician's decision to place a chest tube.
RESULTS: Of the 240 patients with PPE who entered the study, 85 had UPPE, 67 had CPPE, and 88 had empyema. Individual pleural fluid parameters, namely a pH<7.20, a glucose<40 mg/dL or <60 mg/dL, a LDH>1000 U/L or a positive culture had a relatively high specificity (from 78% for LDH to 94% for glucose<40 mg/dL), but low to moderate sensitivity (from 25% for culture to 73% for LDH) in predicting the need for chest tube placement in non-purulent PPE. While pleural fluid cultures performed poorly in discriminating UPPE from CPPE (likelihood ratio positive 1.7), effusion's size performed the best (likelihood ratio positive 5.7). BTS and ACCP guidelines yielded measures of sensitivity (98% and 97%, respectively), and negative likelihood ratio (0.03 and 0.05, respectively) for identifying a CPPE.
CONCLUSIONS: Both guidelines have similar accuracy and perform satisfactorily in distinguishing CPPE from UPPE, albeit at an admissible cost of needlessly increasing chest tube drainage.
PMID 16253494 Respir Med. 2006 May;100(5):933-7. doi: 10.1016/j.rmed.2005.06.017. Epub 2005 Oct 25.
|
