Abstract
OBJECTIVE
To assess the sensitivity and specificity of polymerase chain reaction (PCR) in infants and children at risk for human immunodeficiency virus (HIV) infection.
DESIGN
A prospective, blinded study of 286 HIV-seropositive infants and children. Infection was diagnosed by antibody detection after 18 months of age, two positive direct tests (p24 antigen and HIV culture), or the presence of an illness that defines the acquired immunodeficiency syndrome.
SETTING
University of South Florida and All Children's Hospital, St Petersburg, inpatient and outpatient centers.
PARTICIPANTS
Two hundred eighty-six infants and children seropositive for HIV who were examined between July 1988 and September 1992.
MAIN OUTCOME MEASURES
Sensitivity, specificity, and predictive values of a commercially available PCR test.
RESULTS
Five hundred sixty-seven PCR tests were performed on samples from 286 seropositive subjects followed up for a minimum of 16 months. Of the subjects, 105 were confirmed to be infected and 181 uninfected. Overall, 96 of 105 initial PCRs in infected subjects were positive (sensitivity, 91.4%; positive predictive value, 99%). If samples obtained during the first week of life are excluded, 95 to 100 samples were positive (sensitivity, 95%). Of 181 initial PCR tests from seropositive subjects who seroreverted, 180 were negative (specificity, 99.4%,; negative predictive value, 95.2%). The predictive value of a positive test was 90.9% and that of a negative test was 93.1% in the first month of life. All 145 initial samples obtained between 5 weeks and 12 months of age correctly predicted infection status (positive predictive value, 100%).
CONCLUSIONS
Gene amplification by means of a commercially available PCR is useful in the diagnosis of HIV infection for infants born to seropositive mothers. Between day 7 through 1 year of age, HIV infection is accurately diagnosed by the PCR assay.
Collapse