Quantitative determination of glucose in serum by isotope dilution mass spectrometry following gas liquid chromatography with fused silica column

To evaluate the accuracy of glucose methods, we have set up an isotope dilution gas chromatographic mass spectrometric method where methyloxime trimethylsilyl derivatives of monosaccharides were separated with a SP‐2100 fused silica column. An autodiluter was found suitable for delivering aliquots for derivatization: CV of 0.13% (aqueous) and 0.24% (serum). Ions at m / z 319 (non‐labelled) and 323 (deuterium‐labelled) were selected for glucose, galactose and mannose, and at m / z 307 and 311 for fructose. Optimization of ion monitoring with a Finnigan 4000 gas chromatograph mass spectrometer 6110 data system was limited to one scan per second without full compensation for 0.1 u shifts. A CV of 0.65% in isotope ratios was obtained for 30 injections (three per vial) of one glucose standard. A similar variation between duplicate injections was obtained with serum. Glucose measurements on 36 serum specimens (six individuals) were made by gas chromatography mass spectrometry and by both a manual (reference) and a flow‐through method using hexokinase and glucose‐6‐phosphate dehydrogenase. Mean glucose by gas chromatography mass spectrometry was 5.57 mmol l −1 ; bias was −1.7% manually and +1.2% by the automated method. Average concentrations (mmol l −1 ) of other hexoses were low: mannose (0.058), fructose (0.015) and galactose (0.005). Glucose measurements of two control sera bracketed with standards showed satisfactory agreeement between the three methods. CV of 1.2 to 1.4% by gas chromatography mass spectrometry indicate that further refinements are required.