The aim of the study was to assess the relative influence of mean blood glucose (MBG), glucose instability (GI) and biological variation of glycohemoglobin (BVG) on HbA1c. The study included 378 unselected young type 1 diabetic patients with a diabetes duration > 1 year. There were 1,409 visits with simultaneous HbA1c determinations and self-monitoring of BG meter downloads. GI was quantified by measuring the standard deviation (SD) of the recorded BG values. A statistical model was developed to predict HbA1c from MBG. Hemoglobin glycation index (HGI) was calculated (HGI = observed HbA1c--predicted HbA1c) for each visit to assess BVG based on the directional deviation of observed HbA1c from that predicted by MBG in the model. Afterwards, the population was divided by thirds into high-, moderate-, and low-HGI groups, i.e. high-, moderate-, and low-glycators, reflecting BVG. A total of 246,000 preprandial BG measurements were analysed, with a mean of 177 per visit. Grand MBG +/- SD was 171 +/- 40 mg/dl. Predicted HbA1c was calculated from the equation: 3.8399 + 0.0242 x MBG (r = 0.66; p < 0.0001). A MBG change of 40 mg/dl corresponded to 1% change in HbA1c, within the range 6-12%. Multiple regression analysis showed no significant relationship between SD and HbA1c, after adjustment for MBG. MBG was 10 times more important than SD to predict HbA1c. MBG was not statistically different between the high- and low glycators, but HbA1c was significantly different. Multiple linear regression was used to predict HbA1c from MBG, SD and BVG (measured by HGI), adjusted for age, duration, gender and ethnic origin. BVG and MBG had large influences on HbA1c, the impact of BVG being 84% of the impact of MBG. On the other hand, GI had only 17% of the impact of MBG. In conclusion the effect of BVG on HbA1c is independent and much greater that the influence attributable to GI. Hemoglobin glycation phenotype, responsible for BVG, may be important for the clinical assessment of diabetic patients in order to avoid complications.