Data mining: The association of 2‐h postprandial plasma glucose with the fasting plasma glucose in a large Chinese population

Abstract Background It is generally believed that the lower limit of postprandial plasma glucose is the same or higher than that of fasting plasma glucose (FPG). This study aimed to investigate the relationship between 2‐h postprandial plasma glucose (2‐hPG) and FPG. Insulin sensitivity and β‐cell function were also evaluated. Methods Analytical data from January 2013 to August 2018 included 10 465 participants’ 2‐h OGTT results and 19 518 participants’ FPG and 2‐hPG values after autonomous self‐feeding. Participants were divided into two groups based on the relationship between FPG and 2‐hPG (OGTT‐A1/Postprandial‐B1:FPG > 2‐hPG;OGTT‐A2/Postprandial‐B2:FPG ≤ 2‐hPG).Insulin sensitivity was evaluated by Matsuda index and homeostasis model assessment of insulin resistance (HOMA‐IR). β‐cell function was estimated by homeostasis model assessment of β‐cell function (HOMA‐β) and early‐phase insulin secretion index (ΔI30/ΔG30). Results The ratio of OGTT‐A1 and OGTT‐A2 is 11.1%; the ratio of postprandial B1 and postprandial B2 is 13.7%. HOMA‐IR and HOMA‐β values were lower, while Matsuda index and ΔI30/ΔG30 values were higher in the non‐diabetic OGTT‐A1 group than those in the OGTT‐A2 group. The value of Matsuda index in women was 0.368 times higher than that in men in group OGTT‐A1. In group OGTT‐A2, the values of HOMA‐IR (0.346), HOMA‐β (9.096), and ΔI30/ΔG30 (3.575) in women were lower, higher, and higher than those in men, respectively. Both HOMA‐β and ΔI30/ΔG30 decreased with age in OGTT groups. Conclusion It existed that FPG was >2‐hPG, and this group had better insulin sensitive and β‐cell function. The influence of age on insulin sensitivity and β‐cell function was greater than that of gender.


| INTRODUC TI ON
Diabetes mellitus (DM) is a global epidemic. Along with population aging, urbanization, positive family history, and obesity crowdsourcing, the prevalence of diabetes continues to increase, especially in developing countries. A cross-sectional survey that was conducted in 2013 in China revealed that China has 10.9% overall prevalence of diabetes and 35.7% prediabetic stage. 1 China has become the country with the highest population of people with diabetes in the world. 2 Diabetes is a common metabolic disorder characterized by hyperglycemia. Normally, the maintenance of blood glucose homeostasis depends on hormonal regulation and neuromodulation.
However, hyperglycemia may result, when genetic and environmental factors contribute to the disorder of hormonal regulation and neuromodulation. Previous studies have shown a hyperbolic relationship of hyperglycemia with insulin resistance and beta-cell dysfunction. 3,4 Several surrogate indices using glucose and insulin levels have been devised as alternative measures of insulin sensitivity, including the homeostasis model assessment of insulin resistance (HOMA-IR) and Matsuda index. 5,6 HOMA-IR is a model that incorporates both fasting insulin and glucose levels. 7 Matsuda index is a model that uses dynamic glucose and insulin values obtained during oral glucose tolerance tests (OGTT). 8 The fasting plasma glucose (FPG) and 2-h postprandial plasma glucose (2-hPG) levels are of great significance in the diagnosis of DM. 9 Correct interpretation of the dynamic changes of plasma glucose is very important for clinical diagnosis and treatment. As we know, the normal FPG level is 3.9-6.1 mmol/L. After glucose loading, the plasma glucose rises, reaching its peak in about 30-60 minutes and then declines; it approaches the baseline level in 2 h (2-h postprandial plasma glucose < 7.8 mmol/L).
What is the relationship between FPG and 2-hPG in clinical practice? Must 2-hPG be greater than FPG? To our knowledge, there are no reported studies examining this relationship using large clinical data. In this study, we investigated the relationship between FPG and 2-hPG, in a large Chinese population. In addition, we assessed the insulin sensitivity and beta-cell function in different groups.

| Data collection
There are two research cohorts in this study: In the first research cohort, analysis was based on the 2-h OGTT data and the complete insulin data between January 2013 and August 2018 from Peking Union Medical College Hospital. Data were excluded based on the following criteria: (a) age < 18 years; (b) missing data on sex or age; (c) FPG < 3.9 mmol/L; (d) serum insulin levels > 300 μL U/mL at any of the OGTT points; (e) 30 minutes postprandial plasma glucose was no more than the 0 minutes postprandial plasma glucose; and (f) 30 minutes serum insulin was no more than the 0 minutes serum insulin. Overall, 10 465 participants' 2-h OGTT data were included in this study ( Figure 1).
In the second research cohort, analysis was based on FPG value and 2-hPG value with autonomous self-feeding between January 2013 and August 2018 from Peking Union Medical College Hospital.
Data were excluded based on the following: (a) age < 18 years; (b) missing data on sex or age; and (c) FPG < 3.9 mmol/L. Overall, 19 518 participants' FPG value and 2-hPG value after self-feeding autonomously were included in this study ( Figure 1).

| OGTT
After an overnight fast, participants were given a solution containing the standard 75 g glucose, and venous blood samples were drawn at 0, 30, 60, 90, and 120 minutes for measuring plasma glucose and serum insulin. According to the World Health Organization criteria, diabetes was diagnosed at FPG ≥ 7.0 mmol/L and/or 2-hPG ≥ 11.1 mmol/ L. 10,11

| Laboratory determinations
Plasma glucose was assessed by enzymatic hexokinase photometric assay using Beckman AU2700 analyzer (Beckman Coulter). Serum insulin was assessed using Siemens ADIVA Centaur XP chemiluminescence immunoassay analyzer (Siemens Healthcare Diagnostics Inc). All analyses were standardized.

| Calculations
The trapezoidal rule was used to calculate glucose and insulin areas under the curve during the OGTT. 12

| Statistical analysis
The Kolmogorov-Smirnov test was used to estimate data distribution. Variables with a skewed distribution were presented as medians (interquartile ranges). The Mann-Whitney U test was used to determine the significance between groups. Chi-squared test was used to compare the counting data between groups. Linear regression analysis was used to explore the effect of sex and age on insulin sensitivity and β-cell function. The quoted P values were two-sided, and a P value < .05 was considered statistically significant. All statistical analyses were performed using SPSS 20.0 (SPSS Inc).

| The basic information of the involved participants
As shown in Table 1, the study included a total of 10 465 participants' 2-h OGTT data. Of this, OGTT-A1 and OGTT-A2 groups had 1047 and 9418 participants' data, respectively. The ratio of  From Table 3 and Figure 3, there are 57.7% participants with the highest insulin level at 60 minutes in the OGTT-A1 group but most people reached the peak insulin level at 120 minutes in the OGTT-A2 (41.7%).

| The insulin sensitivity and β-cell function in the DM and non-DM groups
To further assess the difference in insulin sensitivity and β-cell function between the OGTT-A1 and OGTT-A2 groups, the participants were grouped according to whether they had diabetes or not. As shown in Table 4, in diabetic groups, there were more people in OGTT-A2 than in OGTT-A1. The ratio of OGTT-A1 and OGTT-A2 was 1.9%; however, in non-diabetic groups, the ratio of OGTT-A1 and OGTT-A2 was 14.1%. The diabetic patients in both OGTT-A1 and OGTT-A2 groups had higher plasma glucose levels at the four time points than for non-diabetic patients.

| Effect of sex and age on insulin sensitivity and β-cell function
The results in Table 4 showed that there existed significant difference of HOMA-β and ΔI30/ΔG30 between OGTT-A1 and OGTT-A2 groups (all P < .001). In the Table 1, it is easy to find that the sex and age are different between OGTT-A1 and OGTT-A2 groups (Both < 0.05). Therefore, linear regression analysis was used to explore whether the sex and age may have effect on insulin sensitivity and β-cell function. The results were shown in Table S1.  The difference between OGTT-A1 and OGTT-A2. According to the standard beta coefficient in Table S1, the influence of age on HOMA-β, HOMA-IR, ΔI30/ΔG30, and Matsuda index was greater than that of gender. ΔI30/ΔG30 is a good indicator of early secretory function. 16 We also evaluated the function of islet β cells by HOMA-β and ΔI30/ΔG30.

| D ISCUSS I ON
The results showed a lower HOMA-β values in the OGTT-A1 group than those in the OGTT-A2 group. This finding might be because the OGTT-A2 group had stronger insulin resistance, and thus, the islet β cells function was compensating. The value of ΔI30/ΔG30 was higher in the OGTT-A1 group than in the OGTT-A2 group, indicating that the islet β cell function of the OGTT-A1 group was stronger than that in the OGTT-A2 group. *P < .001, the insulin composition ratio difference between OGTT-A1 and OGTT-A2.
TA B L E 3 Insulin composition ratio peaked at different time points in the OGTT experiment

F I G U R E 3 Composition ratio of insulin peaked at different time points in the OGTT experiment
Overall, there existed the situation that the participants with FPG value > 2-hPG value, although the ratio was small, it was not the wrong results and the insulin resistance of the OGTT-A1 group was weaker than that of the OGTT-A2 group, and the β-cell function of the OGTT-A1 group was stronger than that of the OGTT-A2 group.
Therefore, clinicians should pay full attention to these two situations.
The conclusion of our study may provide a reference value for monitoring the progress of the patient's condition and make sure that Researcher lay less emphasis on the effect of gender and age for insulin resistance and beta-cell function. 18 From the Table S1, It is easy to find that women's insulin resistance is weaker than that of men, and women's islet B cells function is better than men. The difference between OGTT-A1 and OGTT-A2 in the non-DM. c The difference between DM and non-DM.
hope to collaborate with foreign researchers to conduct this study worldwide. However, the finding of this study clearly showed a difference in insulin resistance and β-cell function between the OGTT-A1 and OGTT-A2 groups; therefore, these limitations did not affect our conclusions.

| CON CLUS IONS
Findings showed the ratio of OGTT-A1 and OGTT-A 2 is 11.1% suggest that the number of FPG > 2-hPG level was less than the FPG value was far less than those in whom the 2-hPG level was greater than the FPG after OGTT. Overall, the insulin resistance of the OGTT-A1 group was weaker than that of the OGTT-A2 group, and the β-cell function of the OGTT-A1 group was stronger than that of the OGTT-A2 group. Therefore, clinicians should pay full attention to these two situations both when monitoring the progression of the patient's condition and when they need to take appropriate treatment measures.

ACK N OWLED G M ENTS
We express our gratitude to Beckman Coulter (China) Co., Ltd., and Siemens (China) Co., Ltd., for providing technical support and all individuals, primary care doctors, and nurses who participated in the study.

CO N FLI C T O F I NTE R E S T
The authors declare that they have no competing interests.