Using a novel minimally invasive (≤ 1.4 mm) technique to sample minuscule (0.5 microliter) amounts of dermal interstitial fluid (ISF), we assessed the accuracy of its glucose concentrations in predicting concurrently measured venous plasma and capillary plasma glucose concentrations.


A total of 67 adult (37 male and 30 female) volunteers (57 with and 10 without diabetes) with venous plasma glucose levels from 1.6 to 28.4 mmol/l underwent forearm ISF, antecubetal venous, and fingertip capillary sampling.


Rank correlations were 0.974 for ISF 1 vs. 2, 0.954 for ISF vs. venous, 0.935 for ISF vs. capillary, and 0.987 for venous vs. capillary. Median absolute differences were 0.53 mmol/l for ISF 1 vs. 2, 1.33 mmol/l for ISF vs. venous, 1.06 mmol/l for ISF vs. capillary, and 0.56 mmol/l for capillary vs. venous. Equations expressing ISF glucose as a function of venous and capillary glucose and equations expressing capillary glucose as a function of venous glucose had slopes of 0.995, 0.936, and 1.021, respectively (none significantly different from unity), and intercepts of 1.03 mmol/l (P = 0.024), 0.94 mmol/l (P = 0.131), and 0.56 mmol/l (P = 0.041), respectively. Error grid analysis of ISF vs. venous glucose and of capillary vs. venous glucose showed that 97% of the measurements fell within grids A and B.


Dermal ISF sampling is a bloodless minimally invasive technique that provides a medium for glucose measurement, the concentrations of which closely reflect ambient glycemia to a degree comparable with that of capillary glucose measurements.

This content is only available via PDF.