Abstract. We studied the cross-sectional relationship between GA and HbA1c in 142 type 2 diabetic patients who had an HbA1c level <7.5% for at least one year without fluctuation by more than 0.5%. We also followed the changes of GA and HbA1c in 18 type 2 diabetic patients for 16 weeks as they progressed from untreated severe hyperglycemia (HbA1c≥9.0%) to good glycemic control (HbA1c≤6.5%) by intensive insulin treatment. The annual mean levels of GA and HbA1c in the stably controlled patients showed a weak, but significant, correlation (r = 0.23, p<0.001) in the 142 diabetic patients. However, the GA/HbA1c ratio ranged widely from 2.0 to 4.0 showing a normal distribution (2.9 ± 0.34, M ± SE), although patients with conditions affecting albumin turnover or RBC lifespan were excluded from the study. The GA/HbA1c ratio was significantly higher when patients were in hyperglycemic than when glycemic control was good (3.5 ± 0.15 vs. 2.9 ± 0.07, M ± SE, p<0.01). GA decreased more rapidly than HbA1c during intensive insulin therapy, but the percent reduction of HbA1c eventually corresponded with that of GA by 16 weeks after the start of treatment. These results demonstrate that, although unknown influences on GA or HbA1c may exist, GA may be a useful marker for monitoring short-term variations of glycemic control during treatment of diabetic patients.