New approaches to the rapid diagnosis of tuberculous meningitis.

Abstract

The diagnosis of tuberculous meningitis depends on identifying Mycobacterium tuberculosis in the CSF by direct staining or culture. Because of the catastrophic nature of this illness and because effective specific therapy is available, clinicians generally cannot delay treatment to receive a definitive laboratory diagnosis, and therapy is often begun on a presumptive basis. At the same time, other forms of meningitis may mimic tuberculosis, and some of these other forms demand different kinds of specific therapy. Thus, there remains a major need for additional methods of rapidly diagnosing tuberculous meningitis. The reports that appear elsewhere in this issue of the Journal (by Ribera and his colleagues [1], of the measurement of adenosine deaminase activity in CSF, and by Kadival and his co-workers [2], of the detection of mycobacterial antigen in CSF) are important contributions that may lead to new methods for rapidly diagnosing tuberculous meningitis. If mycobacteria were easily identifiable by direct smear of CSF from most patients, the diagnostic problem would not exist. In their recent review of tuberculous meningitis, Molavi and LeFrock [3] stated that mycobacteria were identified by smear of CSF in KWo^O^o of cases in most series and by culture in 45(7o-90(7o of cases. A careful modern study of the bacteriologic diagnosis of tuberculous meningitis is that of Kennedy and Fallon [4]. They found that the initial CSF specimen was positive in 37^0 of patients on direct smear and in 52^0 on culture. Examining second and third CSF specimens added substantial numbers of positives, but examining additional specimens yielded few additional diagnoses. Ultimately, 87^0 of patients had a positive CSF smear and 83 ^o, a positive culture. The diagnostic yield of CSF smear and culture in other studies has generally been lower. Any proposed test for the rapid diagnosis of tuberculous meningitis must be examined, with respect