Glycosylated hemoglobin in relation to rapid fluctuations in blood glucose in children with insulin-dependent diabetes: a comparison of methods with and without prior dialysis.

To evaluate the importance of dialysis in the determination of glycosylated hemoglobin (HbA1), we studied blood glucose and HbA1 in 38 insulin-dependent diabetic children during a morning fast and again 6 h postprandially. We used two methods to determine glycosylated hemoglobin: (1) the conventional macrocolumn method of Trivelli, which uses dialyzed hemolysate and (2) a commercially available microcolumn procedure, Isolab's Fast Hemoglobin Test System, which uses undialyzed blood samples. When the 6-h changes were assessed, the mean blood glucose had increased from 11.6 to 16.3 mmol/L (P less than 0.001). HbA1, determined by the microcolumn procedure simultaneously increased from 12.6% to 13.4% (P less than 0.001), and the increment in HbA1 correlated significantly with the increment in blood glucose (r = 0.62, P less than 0.001). HbA1 determined by the macrocolumn method increased slightly from 13.1% to 13.4% (P less than 0.01), and no correlation was present between the increment in blood glucose and HbA1 (r = -0.02, NS). When the microcolumn procedure was modified by employing dialyzed hemolysate, this method became unaffected by acute blood glucose variations. Therefore, dialysis in sample preparation appears to be important in minimizing the effect of acute changes in blood glucose on the level of glycohemoglobin. Methods in which dialyzed hemolysates are used may be more useful as an index of long-term glucose control.