Diagnostic accuracy of HIV-associated central nervous system toxoplasmosis.

Our objective was to examine the accuracy of diagnosis of HIV-associated central nervous system (CNS) toxoplasmosis. Individuals diagnosed with HIV-associated CNS toxoplasmosis and controls were ascertained from a population-based database. Diagnosis was confirmed by response to therapy or by histology. Symptoms, results of anti-Toxoplasma serology and use of Pneumocystis carinii pneumonia (PCP) prophylaxis were recorded. Central nervous system toxoplasmosis was confirmed in 54 (76%) of 75 patients. Reactive anti-Toxoplasma serology was associated with CNS toxoplasmosis (OR=20.4, 95% CI 3.1-175.8). Adjusting for CD4 and use of dapsone or aerosolized pentamidine, trimethoprim-sulphamethoxazole (TMP-SMX) for PCP prophylaxis was associated with lower likelihood of CNS toxoplasmosis (OR 0.3, 95% CI 0.1-0.7). Diagnosis of CNS toxoplasmosis is often incorrect. Another diagnosis is most likely in patients who are anti-Toxoplasma seronegative or who are receiving prophylactic TMP-SMX.