Roy PM, Colombet I, Durieux P, et al. Systemic review and meta-analysis of strategies for the diagnosis of suspected pulmonary embolism. BMJ.2005;331:259.
Background: Despite technological advances, the diagnosis of pulmonary embolism remains challenging. A large number of diagnostic tests and strategies have been evaluated and yet the test characteristics of each and their practical use remain unclear.
Methods: Pierre-Marie Roy, MD and colleagues carried out a systematic review and meta-analysis to define the likelihood ratios (LRs) for different diagnostic modalities for pulmonary embolism and provide a simple, evidence-based diagnostic algorithm.
The authors performed a literature search from 1990-2003 identifying all articles that evaluated tests or strategies aimed at diagnosing pulmonary embolism. They only selected papers which were prospective, in which participants were recruited consecutively, and which pulmonary angiography was the reference standard for strategies to confirm pulmonary embolism and clinical follow-up or angiography were used for exclusion strategies.
Results: Forty-eight articles (11,004 patients) met the inclusion criteria and examined ventilation/perfusion (V/Q) lung scanning, computed tomography (CT) angiography, leg vein ultrasound (U/S), echocardiography, magnetic resonance (MR) angiography, and the D-dimer test. For the studies done to evaluate tests to confirm the diagnosis of pulmonary embolism, pooled positive likelihood ratios (+LRs) were calculated and were:
click for large version
Note: Diagnostic tests with LRs >10 are excellent tests for confirming the diagnosis.
click for large version
Note: Diagnostic tests with –LR <0.10 are excellent tests for excluding the diagnosis.
For the studies evaluating tests to exclude the diagnosis of pulmonary embolism, pooled negative likelihood ratios (-LR) were calculated and were:
Discussion: With the pooled positive and negative LRs, Roy and colleagues created a diagnostic algorithm, based on initial pretest probabilities, to help “rule in” and “rule out” the diagnosis of pulmonary embolism. Consistent with prior studies, a calculated post-test probability of >85% confirmed the diagnosis while a post-test probability <5% excluded PE.
In patients with a low or moderate pretest probability, pulmonary embolism is adequately excluded in patients with either 1) negative D-dimers or 2) a normal V/Q scan or 3) a negative CT angiogram in combination with a normal venous ultrasound. In patients with moderate or high pre-test probability, pulmonary embolism is confirmed by either 1) a high-probability V/Q scan or 2) a positive CT angiogram or 3) a positive venous ultrasound. Low-probability V/Q scanning, CT angiogram alone, and MR angiography have higher negative likelihood ratios and can only exclude PE in patients with low pre-test probability.
Many hospitalists are using CT angiography as their sole diagnostic test for pulmonary embolism. Based on the systematic review and meta-analysis by Roy and colleagues, we should proceed with caution as, in some patient populations, a positive or negative “spiral CT” does not adequately confirm or exclude the diagnosis of pulmonary embolism. For those that employ V/Q scanning, MR angiography, or D-dimers, the study also helps define how best to use these tests.
Safdar N, Maki DG. Risk of catheter-related bloodstream infection with peripherally inserted central venous catheters used in hospitalized patients. Chest. 2005;128:489.
Background: In recent years, peripherally inserted central catheters (PICCs) have become more popular, initially for long-term outpatient intravenous therapy but also for inpatient venous access. Traditionally, it was assumed that PICC lines have a lower rate of catheter-related bloodstream infection than conventional central venous catheters (CVCs) placed in the internal jugular, subclavian, or femoral veins.