Accuracy of prenatal determination of RhD type status by polymerase chain reaction with amniotic cells.

OBJECTIVE

Our purpose was to determine the accuracy of RhD typing by use of amniocytes obtained at amniocentesis in RhD-negative women.

STUDY

DESIGN

One hundred thirty-five RhD-negative women undergoing amniocentesis for management of suspected alloimmunization (n = 95) or routine second-trimester cytogenetic indications (n = 40) were studied. Amniocytes were then used as template to amplify specific portions of the Rh D and Rh CcEe genes by polymerase chain reaction. The fetal RhD type was confirmed by serologic techniques either after fetal blood sampling or cord blood samples at birth.

RESULTS

Thirty-six fetuses were serologically typed as RhD negative and all 36 were typed RhD negative by polymerase chain reaction. Ninety-eight fetuses were serologically typed as RhD positive; of these, 96 were correctly typed as RhD positive and two were incorrectly typed as RhD negative, with an overall error rate of 1.4%. Both of the errors occurred in a single batch of six samples tested at the same time. In one of these cases the fetus had mild Rh alloimmune disease and required exchange transfusion at birth. In the second case the fetus had severe hydrops fetalis and died in utero at 28 weeks. Deoxyribonucleic acid isolated from fetal blood was tested with the same polymerase chain reaction technique after delivery, and in both cases the fetus was correctly typed as RhD positive. Deoxyribonucleic acid amplification repeatedly failed in one case.

CONCLUSION

Prenatal fetal RhD typing by polymerase chain reaction with amniotic fluid cells is accurate and reliable. In sensitized pregnancies it allows early management of Rh disease and avoids invasive procedures in RhD-negative fetuses. In nonsensitized pregnancies it avoids the use of anti-RhD immunoglobulin after invasive procedures or during pregnancy. To eliminate the possibility of genetic and laboratory sources of errors, we suggest using different sets of primers at two different loci (e.g., exon 4 to 5 and exon 10).