Cotside plasma rather than whole blood estimates of blood glucose in the newborn.

An evaluation of the use of a glucose reflectance meter for the cotside measurement of plasma rather than whole blood glucose was undertaken. Three hundred and twelve samples were obtained for examination. There is a closer correlation between plasma glucose than whole blood glucose to laboratory values. Confidence and prediction intervals show that even separating plasma from whole blood at the cotside does not improve the reliability of glucose reflectance meters when predicting blood glucose values in the context of managing high risk newborn infants. The cotside measurement of blood glucose using glucose reflectance meters should be abandoned and improved methods for quickly obtaining accurate laboratory data should be instituted.