The discovery of at least two cyclooxygenase (COX) isoenzymes had two major consequences: i) to give a new impetus to the research on lipid metabolism, giving rise to the crystallization of these peculiar membrane enzymes, the characterization of their active sites and their gene regulation, and the identification of new metabolic pathways; ii) the development of new NSAIDs aimed to have an improved safety profile, the coxibs. These drugs are defined by their COX-2 selectivity which is supported by a negligible inhibitory potency on platelet COX-1 in vitro and ex vivo after oral intake of maximal therapeutic doses. However, the coxibs marketed in France (celecoxib, rofecoxib, parecoxib) are not equivalent in terms of selectivity and some drugs developed by pharmaceutical companies (etoricoxib, lumiracoxib) will be even more selective for COX-2. These "new" coxibs are the final step in the theory of COX-2 selectivity and they will probably be helpful to better define the limitations of the therapeutic concept based on a selective inhibition of this iso-enzyme.