In emergency, some "low cost" biological tests are too often systematically performed while others are not prescribed because they are considered "too expensive". Good practices for biological testing are in fact the real means for saving money. Here, in order to help for defining those good practices, we review emergency medical approach and, as an example, specify findings concerning the clinical utility of key blood analyses in patients with acute dyspnea. Emergency laboratory testing is usefull when it contributes to establish the diagnosis or to evaluate comorbidity or to stratify disease severity. In a given emergency context, clinical utility must be anticipated according to a bayesian approach with an estimation of the post-test probabilities using the likelihood ratios (estimated from literature) and the pretest probabilities (established by examination at the bedside). The likelihood ratio is the best criterion for diagnostic accuracy of a biological test. According to this criterion, troponin, natriuretic peptides, procalcitonin and D-dimers are four "costly" markers but which can significantly contribute to the etiologic diagnosis of an acute dyspnea. Troponin, natriuretic peptides and procalcitonin are also prognostic markers and are valuable parameters for stratifying disease severity according to their initial value and their plasmatic kinetic during the clinical course of the disease. In conclusion, it is not only the cost of the test but overall the potential impact of its result on the management of the patient's care which makes the decision of performing the test or not.