Diagnosis of myasthenia gravis in the 1990s.

We have tried to show through this analysis how an understanding of the characteristics of diagnostic tests can lead one to a rational strategy for making clinical decisions. The diagnosis of MG can be made with a high degree of confidence in most situations, but the amount of supporting test information necessary depends on the clinical situation. This analysis does show some general guiding principles. The first is that the results of the edrophonium test should be used to guide further diagnostic assessment in most situations. In the small number of cases in which an objective edrophonium test cannot be done, the pretesting estimate of certainty should be used to guide testing strategy. A second principle is that, if a second diagnostic test is required after edrophonium testing, it should be the acetylcholine receptor antibody assay. This test is less expensive and it has more overall reliability. EMG testing thus serves as a third-line test for MG, and it can be reserved for a small number of clinical situations in which certainty about the diagnostic test is too high to reject the possibility of MG, yet too low to initiate treatment confidently. The greatest attractiveness of this type of analysis is that one may take issue with one or more of the assumptions on which it is based without invalidating the conclusions. The conclusions can be tested by a process called sensitivity analysis, in which one or more of the variables can be changed to study its effect on the outcome of the analysis. For example, one could require that certainty exceed 99% (T1) prior to beginning treatment.(ABSTRACT TRUNCATED AT 250 WORDS)