Use of a clinical laboratory database to estimate Toxoplasma seroprevalence among human immunodeficiency virus-infected patients. Overcoming bias in secondary analysis of clinical records.

Abstract

Knowledge of Toxoplasma seroprevalence among human immunodeficiency virus-infected patients identifies candidates for potential prophylaxis and may influence management among patients who present with mass lesions of the central nervous system. Toxoplasma antibody seroprevalence among persons with human immunodeficiency virus infection varies widely. We utilized an administrative and laboratory database (including 1599 patients with human immunodeficiency virus infection who were receiving medical care at the University of California-San Diego Medical Center, between July 1, 1986, and March 31, 1990) to estimate Toxoplasma seroprevalence. Of the 1599 patients, 344 underwent serum assays for Toxoplasma antibodies. The tested group was similar to other members of the cohort in distributions of sex, race, citizenship, and age. The overall mean (+/- SE) Toxoplasma seroprevalence was 16% +/- 2%. Potential bias in this estimate because of testing indication, misclassification, or testing frequency was small.