Altitudes and Hemoglobin A_1c Values: An Analysis Based on Two Nationwide Cross-sectional Studies

People living at high altitudes require an increase in the number of erythrocytes, known as polycythemia, to compensate for decreased oxygen availability (1). Using HbA_1c to detect prediabetes or diabetes at high altitudes might lead to misdiagnosis due to a pseudo-increase in HbA_1c levels in people with polycythemia. In the current study, we aimed to present the effect of high altitudes on diagnosing prediabetes and diabetes using HbA_1c.

The 2010 China Noncommunicable Disease Surveillance study is a nationwide survey that used a complex, multistage, cluster sampling method to recruit participants from 162 study sites across mainland China (2). In total, 95,052 adults aged ≥18 years who were not taking antidiabetes medications were included in the analysis. The prevalences of prediabetes and diabetes at different altitudes were compared using χ² test that considered weights and study design. The diagnostic accuracy of HbA_1c for prediabetes and diabetes, defined by fasting plasma glucose (FPG) and 2-h postload glucose (2-h PG), was calculated separately in participants living below and above 2,500 m of sea level. The optimal cutoff was determined by selecting the point maximizing the Youden index. The areas under the curves (AUCs) were compared to assess the discriminatory ability of HbA_1c by altitude status. To assess the detection bias of HbA_1c caused by high altitudes, a subgroup of participants living at low to high altitudes was selected and matched for age, sex, BMI, FPG, and 2-h PG levels. A total of 2,165 participants living at an altitude above 3,750 m were 1:1:1:1 matched with the participants living below 1,250 m, between 1,250 and 2,499 m, and between 2,500 and 3,749 m. Nearest-neighbor matching was used with the caliper set at 0.05. All analyses were performed using R version 4.2.3. A two-sided P value of <0.05 was considered statistically significant.

There were 88,577 participants living below 2,500 m and 6,475 above 2,500 m. In almost all the categories of FPG (per 5-mg/dL increase) and 2-h PG (per 10-mg/dL increase) levels, mean HbA_1c values were significantly higher in those living above 2,500 m than in those living below 2,500 m. The prevalence of prediabetes and diabetes was significantly higher in the participants living below 2,500 m than in those living above 2,500 m when defined using FPG (32.4% vs. 16.9% for prediabetes; 4.9% vs. 2.2% for diabetes) (Fig. 1A) or 2-h PG (10.7% vs. 5.8% for prediabetes; 3.7% vs. 1.4% for diabetes) (Fig. 1B). In contrast, the prevalence of prediabetes and diabetes was not statistically different among participants living below 2,500 m and above 2,500 m when defined using HbA_1c (38.8% vs. 38.4% for prediabetes; 5.0% vs. 3.6% for diabetes) (Fig. 1C).

The discriminative ability of HbA_1c to identify prediabetes and diabetes defined by FPG and 2-h PG was significantly higher among participants living below 2,500 m versus those living above 2,500 m (AUC difference 3.6%, P < 0.001 for prediabetes; AUC difference 3.9%, P = 0.048 for diabetes) (Fig. 1D–G). The optimal HbA_1c cutoffs to detect prediabetes and diabetes were consistently higher among participants living above 2,500 m than those living below 2,500 m. By using the new data set of matched participants, the associations between altitudes and HbA_1c were flattened below 2,500 m and monotonically increased above 2,500 m (Fig. 1H). We repeated the analyses in another nationwide study population of 48,704 migrant workers aged 18–59 years (3) and obtained similar results (data not shown).

This is the largest study that assessed the impact of altitude on HbA_1c measurements. In line with a previous study (4), we considered that HbA_1c measurements were inaccurate for people living at high altitudes. A proper adjustment might have to be considered when using HbA_1c to diagnose diabetes at an altitude above 2,500 m. Our findings were not only obtained from the long-term dwellers but also applied to people who migrated to their current residence at least 6 months previously (2,3). However, we did not measure hemoglobin concentrations or red blood cell counts, which limits further interpretation. Long-term living in hypoxic environments at high altitudes not only causes an increase in erythrocyte formation but also might affect erythrocyte volume and inhibit eryptosis of mature erythrocytes (5). A longer life span of erythrocytes may prolong the time of hemoglobin exposure to blood glucose and lead to the rise of HbA_1c concentrations.

Overall, HbA_1c may be systematically overestimated for people living above 2,500 m. Therefore, a proper adjustment might have to be considered when using HbA_1c to diagnose diabetes at an altitude above 2,500 m. Further studies are needed to correct the detection bias of HbA_1c caused by high altitudes.