Serological diagnosis of tuberculosis. Evaluation of two years' experience.

Abstract

Two years' experience of a simple serological test for the diagnosis of tuberculosis has been evaluated. This test fails to distinguish clearly between tuberculous and non-tuberculous patients. There is an unacceptably high incidence of false positive and false negative results, although the latter would be reduced by taking the highest titre values of several samples from one patient. Modification of the test to eliminate IgM agglutination produced a marked fall in false positive results. These preliminary results suggest that this modified test is more useful, in that a positive test is a strong indication of tuberculosis. The use of a Coombs' test to agglutinate incomplete complexes did not improve results.