Whole-Blood Glucose Testing at Alternate Sites

Abstract

OBJECTIVE—To measure hematocrit (Hct) and glucose concentration in capillary blood drawn from the fingertip and forearm of a group of 50 nonfasting subjects with diabetes. RESEARCH DESIGN AND METHODS—Hct was determined indirectly by measuring Hb with the HemoCue B-Hemoglobin Photometer. Glucose was assayed with the HemoCue B-Glucose Analyzer, chosen as the independent control assay, and the Sof-Tact Blood Glucose System. Testing session with each subject lasted ∼30 min and consisted of a sequence of tests with each device (SofTact, HemoCue Glucose, and HemoCue Hemoglobin), performed on the arm and then on the fingertip. This sequence was performed three times, so all tests were done in triplicate. Additional fingersticks were performed on each subject at the start and end of the session to measure net change of glucose status during the experiment with a YSI Glucose Analyzer. The mean of the triplicate assays was used to calculate each subject’s percent of glucose difference between arm and finger [(arm glucose − finger glucose)/finger glucose]. Because of the order in which replicates were performed, time-dependent changes in the glucose status of subjects had little effect on the mean values. Thus, the percent of glucose difference calculated herein reflects the intrinsic difference between forearm and fingertip. RESULTS—Hb concentration and Hct were found to be significantly higher in the arm than in the finger. When intraperson differences were calculated, the difference for Hb and Hct was found to be 1.8 ± 1.1 g/dl (mean ± SD) and 5.3 ± 3.0%, respectively. In contrast to Hb, the percent of glucose difference between arm and finger was statistically insignificant. When measured with HemoCue, the percent of glucose difference was −0.1 ± 8% for all 50 subjects, −1 ± 6% for 20 subjects, for whom blood glucose varied <9 mg/dl during the experiment, and 2 ± 10% for 15 subjects, for whom blood glucose varied >18 mg/dl. Thus, irrespective of how much blood glucose changed among the subjects, the glucose difference between forearm and fingertip was insignificant and less than measurement errors. A major source of error in the calculated differences was variability between replicates. No correlation was observed between an individual’s Hct bias and his or her percent of glucose difference, as measured with HemoCue. The results with Sof-Tact were similar, with percent of glucose difference again being statistically insignificant. The measured difference was −4 ± 13% for all 50 subjects, −1 ± 15% for 20 subjects, for whom blood glucose varied <9 mg/dl during the experiment, and −1 ± 12% for 15 subjects, for whom blood glucose varied >18 mg/dl. There was no correlation between a subject’s Hct bias and his or her glucose difference, as measured with Sof-Tact. CONCLUSIONS—In this cross-sectional study of 50 nonfasting subjects whose blood glucose concentration changed to various degrees during the experiment, no significant glucose difference was observed between the capillary beds of the forearm and fingertip, regardless of whether glucose was assayed with HemoCue or the Sof-Tact Blood Glucose System. On the other hand, Hb concentration and Hct were found to be significantly higher in the capillary blood of the forearm.