Abstract
We investigated the association between postchallenge glucose increment and hemoglobin glycation index (HGI), the difference between observed and predicted glycated hemoglobin (HbA1c), in subjects with no history of diabetes. We enrolled 1381 subjects who attended our outpatient clinic for an oral glucose tolerance test (OGTT) to screen for diabetes. HGI was defined as observed HbA1c minus predicted HbA1c. The predicted HbA1c was calculated by entering fasting plasma glucose (FPG) level into an equation [HbA1c(%)=FPG(mg/dL)*0.029+2.9686] determined from an HbA1c versus FPG regression analysis using data from an independent cohort of 2734 subjects with no history of diabetes. The association between 2-hour glucose increment and HGI was analyzed using linear regression analyses with adjustment of relevant parameters. Overall, the proportions of subjects with normal glucose tolerance, pre-diabetes, and newly diagnosed diabetes were 42.3%, 41.3%, and 16.4%, respectively. Compared with subjects who had an HGI≤0, subjects with an HGI>0 had a lower FPG (95.0±13.3 vs 98.5±15.3 mg/dL, p<0.001) but a higher 2-hour plasma glucose (151.1±52.8 vs 144.6±51.4 mg/dL, p=0.027) and 2-hour glucose increment (56.1±46.1 vs 46.1±45.0 mg/dL, p<0.001). The 2-hour glucose increment after an OGTT was independently associated with HGI (β coefficient 0.003, 95% CI 0.002 to 0.003, p<0.001). Our findings suggested that postchallenge glucose increment was independently associated with HGI in subjects with no history of diabetes.
Significance of this study
What is already known about this subject?
It is not unusual to find a considerable discordance between glycated hemoglobin (HbA1c) and blood glucose levels.
The hemoglobin glycation index (HGI, defined as the observed HbA1c minus predicted HbA1c based on glucose levels) has been proposed to quantify those differences (variations).
The mechanisms that lead to an HGI are not yet clear.
What are the new findings?
Compared with subjects who had an HGI≤0, subjects with an HGI>0 had a lower fasting plasma glucose but a higher 2-hour plasma glucose and 2-hour glucose increment.
The findings were observed in those who had abnormal glucose regulation (diabetes or pre-diabetes) determined by an oral glucose tolerance test (OGTT).
The 2-hour glucose increment was independently associated with HGI.
How might these results change the focus of research or clinical practice?
For subjects at risk for diabetes who have a high HGI (eg, a higher HbA1c in relation to fasting glucose), an OGTT is recommended as they are likely to have postchallenge hyperglycemia.
For patients with diabetes with a high HGI (eg, a high HbA1c with a relatively low fasting glucose), monitoring of postprandial glucose is recommended as they are likely to have postprandial hyperglycemia.
HGI may be useful for individualized glycemic therapy in patients with diabetes.
Introduction
Glycated hemoglobin (HbA1c) has been widely used for the assessment of glycemic control in people with diabetes and for the diagnosis of abnormal glucose regulation (diabetes or pre-diabetes) in people with no history of diabetes.1 2 Nevertheless, the correlation between HbA1c and blood glucose levels is only modest.3 4 Therefore, in assessments of glycemic control5 and diagnosis of abnormal glucose regulation,6–8 it is not unusual to find a considerable discordance between HbA1c and blood glucose levels. Such discordances may complicate data interpretation,9 and it is clinically relevant to understand factors that may alter the association between HbA1c and glucose levels.
HbA1c values are influenced by both glycemic and non-glycemic factors.10 There are considerable between-individual differences (variations) in the relationship of HbA1c to glucose levels.4 11 The hemoglobin glycation index (HGI) has been proposed to quantify those differences (variations).12–14 HGI is defined as the observed HbA1c minus predicted HbA1c based on glucose levels. The predicted HbA1c can be derived by using a linear regression between HbA1c and blood glucose levels.11–13 Hence, subjects with a high HGI have an HbA1c level higher than that estimated from their blood glucose levels. As HGI has been associated with cardiovascular diseases15 and mortality,16 it is important to clarify factors that may be associated with HGI.
Although the mechanisms that lead to between-individual variations in the relationship of HbA1c to glucose levels are not yet clear,15 17 postprandial hyperglycemia has been suggested to be associated with increased glycation gap/HGI.17 However, there are few data showing an association between postprandial hyperglycemia and glycation gap/HGI.18 We hypothesized that HGI was associated with postprandial glucose increment (postprandial hyperglycemia). In this study, we investigated the association between postchallenge glucose increment and HGI in subjects with no history of diabetes who underwent an oral glucose tolerance test (OGTT) by using fasting glucose levels to determine the predicted HbA1c levels and HGI.15 19
Materials and methods
We enrolled subjects with no history of diabetes who attended our outpatient clinic for an OGTT to screen for diabetes between 2011 and 2019. We excluded subjects from the analysis if they had blood transfusion within 3 months or anemia, defined as a hemoglobin level less than 100 g/L.
All subjects underwent a 75 g OGTT.20 Briefly, subjects fasted overnight and attended our outpatient clinic in the morning. A trained nurse interviewed the subjects and measured their height, weight, waist circumference, and blood pressure. For the measurements of plasma glucose, insulin, HbA1c, and lipids profiles, we collected a fasting blood sample before the OGTT. We then conducted a 75 g OGTT20 for study subjects, and collected a blood sample after 2 hours to determine the postchallenge plasma glucose levels.
We measured HbA1c by using boronate-affinity high-performance liquid chromatography with an interassay and intra-assay coefficient of variation of <2.9% and <0.9%, respectively. We determined plasma glucose by using the glucose oxidase-peroxidase method. The interassay and intra-assay coefficients of variation were both <1.5%. We measured plasma insulin by using an electrochemiluminescence immunoassay with an interassay and intra-assay coefficient of variation of 2.5% and 1.8%, respectively.
We divided the study population into 2 groups according to their HGI (the observed HbA1c minus predicted HbA1c). We developed an equation to estimate HbA1c by fasting plasma glucose (FPG) in the study population.15 19 The equation was derived using data from an independent cohort (developmental database) of 2734 subjects with no history of diabetes who had visited our hospital for a health check-up (mean age 47.4±11.2 years, male 66.4%, mean body mass index 24.4±3.6 kg/m2, mean FPG 91.7±17.7 mg/dL, mean HbA1c 5.6%±0.6%). We excluded subjects from the developmental database if they had blood transfusion within 3 months or anemia, defined as a hemoglobin level less than 100 g/L. An equation [HbA1c(%)=FPG(mg/dL)*0.029+2.9686] was determined based on the correlation between HbA1c and FPG in the developmental database (figure 1).
We determined glucose regulation status of our study subjects according to OGTT, as recommended by the American Diabetes Association.21 We assessed insulin resistance and β-cell function using the homeostasis model assessment (HOMA-IR and HOMA-β, respectively).22 To investigate the association between postchallenge hyperglycemia and HGI, 2-hour glucose increment (2-hour plasma glucose—FPG) was calculated using OGTT results.
The Statistical Package for the Social Sciences (IBM SPSS V.22.0; International Business Machines, New York, USA) was used to perform the statistical analyses. Categorical and continuous data are expressed as numbers (percentages) and mean±SD, respectively. Kolmogorov-Smirnov test was applied to determine whether a variable was normally distributed. To examine statistical differences in continuous variables between groups, the Student’s t-test was used. For categorical variables, the χ2 test was used. To examine the association between 2-hour glucose increment and HGI, linear regression analyses were used with adjustment of relevant parameters. Bonferroni correction was applied for multiple comparisons. Data that were not normally distributed were logarithmically transformed before analysis. Statistical significance was considered for a two-sided p value <0.05.
Results
A total of 1381 outpatients (mean age 60.5±11.6 years, male 82.3%, mean body mass index 26.1±3.6 kg/m2) underwent an OGTT. Using the equation [HbA1c(%)=FPG(mg/dL)*0.029+2.9686] derived from the developmental database (figure 1) to predict HbA1c and calculate HGI, we divided the study population into 2 groups (HGI≤0 vs HGI>0, table 1). Subjects with an HGI>0 had a lower FPG (95.0±13.3 vs 98.5±15.3 mg/dL, p<0.001) and a lower predicted HbA1c (5.7%±0.4% vs 5.8%±0.4%, p<0.001), but a higher observed HbA1c (6.1%±0.5% vs 5.5%±0.5%, p<0.001), compared with those who had an HGI≤0. Body mass index (26.3±3.6 vs 25.8±3.5 kg/m2, p=0.013), total cholesterol (186±40 vs 180±39 mg/dL, p=0.011), and triglycerides (156±114 vs 142±97 mg/dL, p=0.035) were higher in subjects with an HGI>0 (table 1). No significant difference was observed with regard to the other parameters between the 2 groups.
Overall, the proportions of subjects with normal glucose tolerance, pre-diabetes, and newly diagnosed diabetes were 42.3%, 41.3%, and 16.4%, respectively. The corresponding values in subjects with an HGI≤0 were 42.6%, 42.2%, and 15.2%. In subjects with an HGI>0, these values were 42.1%, 40.7%, and 17.2%. Nevertheless, we observed a lower rate of isolated impaired fasting glucose (4.4% vs 11.9%) and a higher rate of isolated impaired glucose tolerance (25.7% vs 21.2%) in subjects with an HGI>0 (table 1).
Table 2 shows fasting and 2-hour plasma glucose, 2-hour glucose increment, HOMA-IR, and HOMA-β according to HGI. Compared with subjects who had an HGI≤0, subjects with an HGI>0 had a lower FPG but a higher 2-hour plasma glucose and 2-hour glucose increment. The aforementioned findings were observed in those who had abnormal glucose regulation (diabetes or pre-diabetes) by OGTT, but not in those who had normal glucose tolerance. We did not observe significant differences in insulin resistance (HOMA-IR) and β-cell function (HOMA-β) between the 2 groups.
Table 3 shows the association between 2-hour glucose increment and HGI. Using linear regression analysis, we demonstrated that 2-hour glucose increment after an OGTT was positively associated with HGI (β coefficient 0.002, 95% CI 0.002 to 0.003, p<0.001). The association remained significant (β coefficient 0.003, 95% CI 0.002 to 0.003, p<0.001) after adjustment for some parameters, such as age and sex (table 3).
Discussion
In this study, we demonstrated that subjects with no history of diabetes who had an HGI>0 had a higher 2-hour plasma glucose and 2-hour glucose increment after OGTT, compared with those who had an HGI≤0 (table 2). The findings were mainly observed in subjects with diabetes or pre-diabetes defined by an OGTT (table 2). Moreover, 2-hour glucose increment was independently associated with HGI (table 3). Our findings suggest that postchallenge glucose increment may contribute, at least partly, to a higher than predicted HbA1c (HGI>0) in subjects with no history of diabetes.
The mechanisms involved in the ‘glycation gap’ (HGI) are not yet clear. It has been reported that there were ethnic differences in the relationship between glucose concentrations and HbA1c levels.23–25 For a given blood glucose level in patients with type 2 diabetes,23 non-Caucasians had a higher HbA1c level (ie, a high HGI) than Caucasians. Similar findings related to racial differences in the glycation gap were reported in patients with type 1 diabetes24 and impaired glucose tolerance.25 However, race only partially explains the observed glycation gaps.24 In our study, we investigated the association between 2-hour glucose increment and HGI in a Chinese population with no history of diabetes. Our findings suggest that 2-hour glucose increment after OGTT was independently associated with HGI. In a previous study conducted in Europeans with no diabetes,26 there was no significant difference in 2-hour glucose levels among HGI tertiles. Similar finding was noted in Americans at risk for diabetes.27 In contrast, there was a higher 2-hour glucose level and a greater 2-hour glucose increment in the highest HGI tertile in an Asian population with pre-diabetes or diabetes.18 Hence, ethnicity may have some effects on HGI which may help explain our results and previous findings.18 26 27
It is reasonable to speculate that postchallenge hyperglycemia may be associated with a higher than predicted HbA1c. Treatment guidelines suggest treating postprandial glucose excursions in patients with diabetes with an HbA1c above target despite having FPG at target.28 Moreover, treatment targeting fasting or postprandial glucose may have different effects on HGI. In a randomized trial29 which compared the effects of treating prandial versus fasting glycemia on cardiovascular outcomes in patients with type 2 diabetes, patients treated with basal insulin had a lower fasting glucose than those treated with prandial insulin (7.0±0.2 vs 8.1±0.2 mmol/L, p<0.001). However, there was no significant difference in HbA1c between the 2 groups (7.8%±0.1% vs 7.7%±0.1%, p=0.4).29 Thus, the patients would have a lower estimated HbA1c and a higher HGI after treatment with basal insulin, compared with treatment with prandial insulin. Similar findings could be observed in the Treating to Target in Type 2 Diabetes trial.30 In a recent study31 using data from flash glucose monitoring in a Chinese population of patients with type 2 diabetes, the authors reported differences in glycation gap (estimated HbA1c vs measured HbA1c) in subjects with different mean glucose levels. Their findings suggest that different glucose levels may have some effects on glycation gap. It might be possible that glucose regulation status may influence HGI. People with normal glucose regulation may have much smaller difference in glycation gap, or a low glucose peak might be less likely to trigger glycation in vivo. Our results and the aforementioned observations suggest an association between postchallenge glucose increment and HGI.
In addition to ethnicity and postchallenge glucose excursions, several factors were shown to be related to HGI.15 17 18 32 For example, Dunmore et al 33 demonstrated that variations in the level of intracellular deglycating enzyme fructosamine-3-kinase were associated with the glycation gap in human subjects. Liu et al 34 reported an independent association between inflammatory markers and HGI in a population with no history of diabetes. As postprandial hyperglycemia has been associated with peripheral inflammation gene expression,35 this may help explain our finding that 2-hour glucose increment was associated with HGI. Genetic factors32 36 and differences in mean red cell age36 37 had also been suggested to account for the glycation gap. These non-glycemic factors related to HGI may help explain the only modest (though significant) correlation between 2-hour glucose increment and HGI in our subjects (table 3). Furthermore, HGI was found to be associated with chronic diabetes complications,38 cardiovascular disease,15 39 and mortality.16 40 Although the mechanisms underlying the relationships between the aforementioned factors and glycation gap/HGI have yet to be fully elucidated, our findings may have important implications for clinical practice with respect to screening for diabetes and treatment for patients with diabetes. For subjects at risk for diabetes who have a high HGI (eg, a higher HbA1c in relation to fasting glucose), an OGTT is recommended as they are likely to have postchallenge hyperglycemia. For patients with diabetes with a high HGI (eg, a high HbA1c with a relatively low fasting glucose), monitoring of postprandial glucose is recommended as they are likely to have postprandial hyperglycemia. We suggest HGI may be useful for individualized glycemic therapy in patients with diabetes.19 41
There are several limitations in this study. First, this study was conducted in a Chinese population. The association between postchallenge glucose increment and HGI needs to be investigated in other ethnicities. Second, this study investigated subjects at risk for diabetes and thus our findings were not confounded by glucose-lowering therapies. Nevertheless, whether our findings could be generalized to subjects with known diabetes or the general population merits further study. Third, postchallenge hyperglycemia was assessed using an OGTT in this study. Thus, the poor reproducibility of OGTT might confound our results.42 Last, non-glycemic factors which may impact hemoglobin glycation43 were not addressed in our study. With these limitations in mind, the novel association between postchallenge glucose increment and HGI in this study may help identify subjects with postchallenge glucose excursions in daily practice.
In conclusion, we demonstrated that postchallenge glucose increment was independently associated with HGI in subjects with no history of diabetes. Our findings need to be confirmed in other populations.
Data availability statement
Data are available upon reasonable request. Data cannot be made publicly available due to the ‘Personal Information Protection Act’.
Ethics statements
Ethics approval
The study protocol was reviewed and approved by the Institutional Review Boards of Taichung Veterans General Hospital, Taichung, Taiwan (approval number C08215B). This study was conducted in accordance with the Declaration of Helsinki, and written informed consent was provided by all subjects prior to the procedures.
Footnotes
Contributors JSW, WJL, and WHHS designed the research. JSW, ITL, SYL, WLL, and KWL conducted the research and collected the data. JSW and WLL analyzed the data. JSW, ITL, and WHHS wrote the first draft of the manuscript. WJL, SYL, WLL and KWL revised the manuscript critically for important intellectual content. All authors approved the final draft of the manuscript.
Funding This study was supported by Taichung Veterans General Hospital, Taichung, Taiwan (TCVGH-YM1070102 and TCVGH-1073505C, 2018; TCVGH-1083505C, 2019).
Competing interests None declared.
Provenance and peer review Not commissioned; externally peer reviewed.