The smallest amount that an analytical system can detect and measure reliably has been called its “sensitivity” or “limit of detection”. Although many reports have described this performance characteristic, the various protocols that have been applied have been disharmonious. A working group of the Japan Society of Clinical Chemistry (JSCC) was given a mission in 1997 to establish definitions and to develop recommended procedures for this purpose. The working group started in cooperation with the NCCLS because the NCCLS already had a subcommittee sharing this aim. Two performance characteristics have been clearly defined by NCCLS subcommittee. The first is the smallest amount of analyte the analytical method can qualitatively measure in a specimen. This is the limit of detection (LOD). The second is the amount the method can quantitatively measure reliably. This is called the limit of quantitation (LOQ). Together, the LOD and LOQ should be able to identify the capabilities of the laboratory method at the lower end of the measurement scale in terms of clinical units. We have evaluated the NCCLS-recommended procedures in the studies of the existing methods for measuring CRP (C-reactive protein), and found the recommendation acceptable but some difficulties remained in regard to the performance of the procedures. The difficulties were: 1) The matrix effects between the ZERO specimen and those containing the actual amounts of an analyte should be eliminated. 2) A few standards are assigned values in terms of the SI units but others are not. 3) A comparison between two or more methods is difficult if each method's calibrator is not traceable to the primary standard.

View full abstract