Tuberculous meningitis among adults with and without HIV infection. Experience in an urban public hospital.

BACKGROUND

Tuberculous meningitis remains a frequently diagnosed entity in urban US hospitals, with the incidence increasing as a consequence of infection with the human immunodeficiency virus (HIV).

OBJECTIVE

To describe the occurrence, characteristics, and therapeutic responses of tuberculous meningitis among adult patients of an urban public hospital, with special attention to the effects of HIV infection.

DESIGN

Retrospective clinical review of all cases identified among adults over a 12-year interval, collecting demographic and clinical variables.

SETTING

A 550-bed urban teaching hospital.

MAIN OUTCOME MEASURE

Nine-month outcome stratified by survival.

RESULTS

Among 31 adult patients identified as having definite or probable tuberculous meningitis, a majority (n = 20 [65%]) were infected with HIV. Cumulative rates of occurrence per 100 000 persons over the 12 years of the study were estimated at 1.72 for those without HIV infection and 400 for those with HIV infection. The most common symptoms at presentation were fever (83% [24/ 29]) and abnormal mental status (71% [20/28]). One or more abnormalities were present in the cerebrospinal fluid of 97% (30/31) of subjects, and 74% (23/31) had cerebrospinal fluid cultures positive for Mycobacterium tuberculosis. Neuroimaging of 28 patients revealed 1 or more abnormalities in 82% (n = 23). Among 30 patients with available outcome data at 9 months, 43% (n = 13) had died, 40% (n = 12) had survived without sequelae, and 17% (n = 5) had survived with morbidity. HIV infection had no discernible effect on findings.

CONCLUSIONS

Tuberculous meningitis remains relatively common among indigent urban nonwhite populations. While HIV infection has contributed to the increased incidence of tuberculous meningitis, it has not significantly altered the presenting clinical, laboratory, or radiographic findings or the response to therapy of this disease. Parameters associated in a multivariate regression analysis with mortality at 9 months were black race and the absence of corticosteroid use.