PubMed1976
Dextrostix is now widely used as a method of screening for hypoglycemia of the newborn. There has always been some anxiety about its accuracy for estimating very low blood sugars; this is important, since symptomatic hypoglycemia of the newborn does not usually occur until the blood glucose falls below 20 mg/dl. In 1970 a reflectance meter was introduced which would measure the colour of the strip electronically. The aim of this study was to assess the accuracy of the meter and its reliability in estimating hypoglycemia. The study was continued to investigate a new model of the reflectance meter which superseded the original one. In the pilot study using the original model, blood samples were taken from 46 babies. Readings of the Dextrostix were made by two independent observers and compared with a reading taken on the reflectance meter. These estimations were later compared with blood glucose measured by the glucose oxidase method. the meter showed a marked tendency to overestimate the blood glucose: 44 out of 46 samples were overestimated (Fig. 1, 2). In the second series 180 cord blood samples were collected. Because the intention was to study very low levels of blood glucose, the samples were allowed to stand at room temperature for several hours to allow glycolysis to occur. Again, readings were taken by two independent observers and compared with the readings taken on the new Dextrostix-Eyetone meter. The blood glucose was measured on each of the samples. There was a strong correlation (r = 0.8877, p less than 0.00005) between the blood glucose values and the readings taken from the meter, with no tendency towards overestimation (Fig. 3). A similar correlation (r = 0.8533, p less than 0.00005) was seen for the observers' readings and the chemical method, although there was a tendency to underestimate blood glucose (Fig. 4). When the meter gave an estimate of more than 20 ml/dl, in no case was the actual blood glucose in the profoundly hypoglycemic group of less than 10 mg/dl. When estimated by eye there was one case in which the blood glucose was only 7 mg/dl but the observer had estimated the result of 30 mg/dl. In order to avoid any possibility that a blood glucose was less than 20 mg/dl, it is necessary to take action on any estimate below 40 mg/dl on a Dextrostix. Despite this, Dextrostix remains a very useful method of screening for neonatal hypoglycemia, whether assessed by eye or with the new Dextrostix-Eyetone meter. the meter does give better results than estimating Dextrostix by eye, but the difference was never statistically significant.