Comparison of Liquid and Dried Blood for Neonatal Hemoglobinopathy Screening: Laboratory and Programmatic Issues

Abstract

To compare laboratory and programmatic issues in neonatal hemoglobin screening in two systems using either liquid cord blood or heel puncture blood dried on filter paper (DB) to determine the accuracy of hemoglobin phenotype; the collection rate of cord blood versus DB; and the types of and reasons for errors. Cord blood samples for 6904 newborns were analyzed by electrophoresis in a large hospital laboratory. DB samples on the same cohort were analyzed by isoelectric focusing in a state public health laboratory. Interlaboratory concordance was 99.4% for 6904 matched specimen pairs, which included 27 disease and 596 carrier phenotypes. In 42 pairs, discordances for potential disease phenotypes occurred with cord blood samples. Errors involving carrier phenotypes occurred in 15 cord blood samples and 14 DB samples. Inconclusive results requiring repeat testing occurred more often in the cord blood testing system (92) than in the DB system (23). Clerical transcriptions and limitations of the techniques accounted for most laboratory errors. Noncompliance in sample collection occurred more frequently with cord blood (181) than DB (86). Both systems are subject to errors, but are equally reliable for neonatal hemoglobinopathy screening.