[Associations between glycated hemoglobin and glucose indicators in adults in areas at different altitude in China]

Objective: To explore the associations of glycated hemoglobin (HbA1c) with FPG and oral glucose tolerance test 2-hour (OGTT-2 h) in areas at different altitude in China. Methods: Subjects who participated in 2018-2019 China Chronic Disease and Risk Factor Surveillance and had no prior type 2 diabetes diagnosis were included. Subsequently, they were categorized into three groups based on altitude of living area (<2 000, 2 000- and ≥3 000 m). With adjustment for intracluster correlation, multivariable linear regression analysis was performed to evaluate the associations of HbA1c with FPG and OGTT-2 h in the context of HbA1c was normal (<5.7%) or abnormal (≥5.7%). Furthermore, the shape of relationships between HbA1c and glucose indicators was examined using restricted cubic spline. Finally, receiver operating characteristic curve was used to evaluate the diagnostic performance of HbA1c for diabetes. Results: A total of 157 277 subjects were included in the analysis. While FPG and OGTT-2 h levels gradually decreased with increase of altitude, HbA1c level was similar among the three groups. When HbA1c was <5.7%, its association with FPG and OGTT-2 h was weak and no obvious difference was observed among the three groups. When HbA1c was ≥5.7%, the FPG and OGTT-2 h increased by 15.45% (95%CI:14.71%- 16.18%) and 24.54% (95%CI:23.18%-25.91%) respectively per one standard deviation increase in HbA1c in group in area at altitude <2 000 m. However, the FPG and OGTT-2 h increased by 13.08% (95%CI:10.46%-15.76%) and 21.72% (95%CI:16.39%-27.31%), respectively, in group in area at altitude 2 000- m, and increased by 11.41% (95%CI:9.32%-13.53%) and 20.03% (95%CI:15.38%- 24.86%), respectively, in group of altitude ≥3 000 m. The restricted cubic spline indicated that the curve showing the association of HbA1c with FPG and OGTT-2 h was flat when HbA1c was <5.7%, but showed a positive linear relationship when HbA1c was ≥5.7%. The area under curve for detecting diabetes was 0.808 (95%CI:0.803-0.812) in group of altitude <2 000 m and 0.728 (95%CI:0.660-0.796, P=0.022) in group of altitude ≥3 000 m. The relevant optimal cutoff value of HbA1c was 5.7%, with a sensitivity of 65.4% and a specificity of 83.0%, and 6.0%, with a sensitivity of 48.3% and a specificity of 93.7%, respectively. Conclusions: When HbA1c was ≥5.7%, the association between HbA1c and glucose indicators became weaker as the increase of altitude. In the area at altitude ≥3 000 m, it may not be appropriate to use HbA1c in the diagnosis of diabetes.