Prenatal diagnosis using polymerase chain reaction on amniotic fluid for congenital toxoplasmosis

Objective: To evaluate sensitivity, specificity, and predictive values of a prenatal amniotic fluid (AF) polymerase chain reaction (PCR) test for diag nosis of congenital toxoplasmosis. Methods: A multicenter prospective study was done on 271 women with proved primary Toxoplasma infection during pregnancy and who had amniocentesis for prenatal diagnosis by PCR. Live-born infants were eligible for analysis on ly if a serologic follow-up could assess a definitive infection status. Results: Of the 270 evaluable cases, 75 were congenitally infected, 48 of w hom had a positive PCR at prenatal diagnosis. Overall sensitivity of PCR on AF was estimated at 64% (95% confidence interval [CI] 53.1%, 74.9%), negat ive predictive value of 87.8% (95% CI 83.5%, 92.1%), whereas specificity an d positive predictive value were 100% (95% CIs 98%, 100% and 92.3%, 100%, r espectively). Among cases with congenital toxoplasmosis, there were no sign ificant differences between those with positive or negative PCR with regard to median gestational age at maternal infection, interval between maternal infection and amniocentesis, or duration of treatment before amniocentesis . However, sensitivity of PCR was found to be significantly higher for mate rnal infections that occurred between 17 and 21 weeks' gestation (P <.02). Conclusion: A negative PCR of AF cannot rule out congenital infection. In t his case, continuation of treatment with spiramycin combined with ultrasono graphic follow-up and postnatal follow-up are warranted. Our results also s uggest presumptive treatment combining pyrimethamine and sulfonamides in ca se of maternal infection occurring late in pregnancy. (Obstet Gynecol 2001; 97:296-300. (C) 2001 by The American College of Obstetricians and Gynecolog ists.).