Background: The reference interval on a viewpoint of preventive medicine is neccessary for the health testing. The reference interval of HbAlc has been documented in the report of the committee on an interlaboratory standardization of HbAlc determination, Japan Diabetes Society, 1994, and in the preservation law of elderly's health from the Labor Ministry, 1996, individually, and they were different. We sent a questionnaire on the reference interval of HbAlc on April, 2001, to the representative centers of the health testing in Kansai district, which are recognized by japan Society of Multiphasic Testing and Services. The upper value of their reference interval for HbAlc ranged from 5.5 to 6.0%, except for a center which was under consideration about the value.
This study aimed to find the reference interval for HbAlc based on a viewpoint of preventive medicine.
Subjects and Methods: In the annual health checkup for the employees of Shimadzu Corporation working in Kyoto prefecture, those revealing FPG more than 110 mg/d
l or postgrandial glycosuria without FPG measurement have undergone 75 gOGTT for more than 10 years in combination with HbAlc measurement since 1999. Based on the OGTT performed once a year as a rule, subjects were classified into normal type, high normal type, borderline type, diabetes type, and diabetes; those below the degree in which the drug administration is necessary for control of glucose metabolism. Patients with the latter type was arbitrarily defined as those with FPG less than 140 mg/d
l and/or 2 hPG less than 240 mg/d
l. These types were subdivided by those with or without transfer to the other types for 4 years and HbAlc of each subgroup was calculated for its distribution, average, and standard deviation. HbAlc was measured by immunoassay. After the double elimination out of the range of the average±3 SD, CV values of the measurement of two samples of HbAlc; 5.0% and 10.0-11.0% levels in 1999, 2000, and 2001 were 3.7, 3.3, and 2.8% for the former level and 4.7, 6.6, and 3.3% for the latter level, respectively.
Results: 946 man-types and 1460 samples of HbAlc were enrolled. The averages of HbAlc of normal type, high normal type, borderline type, diabetes type, and diabetes were 4.7±0.3, 4.8±0.3, 4.9±0.4, and 5.4±0.4%, respectively. HbAlcs of the other types were significantly higher than those of normal type (p, 0.0005) . The averages of normal type in the successive two OGTTs including the initial one and the other types in the successive 4 OGTTs [A type (-) ] were compared with those of types transferring to the other types [A type (B type) ] . Normal type (borderline type) and normal type (diabetes type) were significantly higher in HbAlc than normal type (-) . High normal type (borderline type, diabetes type, and diabetes) were significantly higher in HbAlc than high normal type (-) and high normal type (normal type) . Borderline type (diabetes type and diabetes) were significantly higher in HbA lc than borderline type (-) and borderline type (normal type and high normal type) . The averages of HbAlc of normal type (-), high normal type (-), borderline type (-), and diabetes (-) were 4.6±0.3, 4.8±0.3, 5.1±0.3, 5.5±0.4%, respectively. HbAlc of the other types (-) were significantly higher than those of normal type (-) (p, 0.0005) .
Conclusions: Patients with impaired glycemia below diabetes stage are vulnerable to change the pathophysiology. In such a situation, the level of HbAlc is usuful for prediction of transferring type. The reference interval of HbAlc percentage calculated from HbAlc results of healthy subjects showing the successive normal OGTT was from 4.0% to 5.2%.
View full abstract