Two major variables may cause false-negative results in prenatal diagnosis of congenital human cytomegalovirus (HCMV) infection: sensitivity of the techniques(s) used; and time elapsed between maternal infection and antenatal testing. Previous results indicated that rapid HCMV isolation from amniotic fluid samples and viral DNA detection in amniotic fluid by nested polymerase chain reaction (nPCR) had comparable levels of sensitivity (69.2% and 76.9%, respectively). The nPCR protocol was reviewed following two additional false-negative antenatal diagnosis in a twin pregnancy during which two procedures were performed at 18 and 23 weeks of gestation, respectively. In the new assay, multiple (instead of single) and 100 (instead of 20) microliters amniotic fluid aliquots were individually amplified and tested by nPCR. By using this approach, low DNA levels (1-10 genome equivalents) were detected in 1-5/8 replicates of amniotic fluid samples taken from both twins during both procedures. In addition, viral DNA was detected in 5/6 replicates from two amniotic fluid samples still available from two previous false-negative cases. However, nPCR on multiple amniotic fluid replicates did not anticipate positive prenatal results in a retrospective case, which required two procedures for correct diagnosis and, when prospectively employed, did not avoid one additional false-negative prenatal diagnosis 8 weeks after maternal infection. Thus, delayed intrauterine transmission of the infection may be a potential cause of false-negative results. However, the combination of a very sensitive technique with appropriate timing of prenatal testing can substantially increase the reliability of prenatal diagnosis results.