In their report, Ericsson and Sherris stated that "adequately described sensitivity categorization schemes have been based on four main concepts" and went on to describe the advantages and inherent limitations of those four criteria. These limitations are very comparable with what is increasingly viewed as limitations to the current breakpoint systems because those four concepts have always been tried to be brought together in one breakpoint system. In line of this development, the European Committee on Antimicrobial Susceptibility Testing has recently recognized that clinical breakpoints based on PK-PD relationships confer a different meaning to resistance than early detection of microorganisms that do not belong to the natural bacterial population, but somehow have acquired resistance mechanisms and have introduced a wild-type cutoff breakpoint conveying a different meaning than clinical breakpoint. Similarly, there is increasing worry on the setting of clinical breakpoints based on frequency distributions because they do not bear an optimal relationship between dose and bacteriologic or clinical effect. This is especially apparent for older drugs because PK-PD was not used until a few years ago. Because dose-effect relationships have now been reasonably well established for most drugs, these should now be used to reappraise current clinical breakpoints.