High one-hour plasma glucose is an intermediate risk state and an early predictor of type 2 diabetes in a longitudinal Korean cohort

Abstract

Aims:

Because one-hour post-load plasma glucose (1h-PG) ≥155 mg/dL (8.6 mmol/L) has been proposed as an early marker for future diabetes but lacks sufficient longitudinal confirmation of its risk, we aimed to evaluate the risk of T2D based on 1h-PG and track changes of insulin sensitivity and β-cell function over time by 1h-PG in a longitudinal cohort.

Methods:

OGTTs were conducted every 2 years in the 10-year longitudinal Korean Genome Epidemiology study (n=6144) with three groups characterized at baseline: Low 1h-PG (< 155 mg/dL) with Normal Glucose Tolerance (NGT), High 1h-PG (≥155 mg/dL) with NGT, and prediabetes (PreDM).

Results:

T2D risk was higher in people with High 1h-PG with NGT and PreDM than those with Low 1h-PG with NGT. Baseline insulin sensitivity in Low 1h-PG as measured by the insulin sensitivity and secretion (ISS) model and Matsuda insulin sensitivity index (ISI) was higher than in High 1h-PG, which was comparable to PreDM. β-cell function as assessed by ISS and the insulinogenic index decreased from Low 1h-PG to High 1h-PG to PreDM. Over time, insulin sensitivity decreased in the three groups. Time from High 1h-PG to T2D was 0.9 years shorter than from Low 1h-PG. All participants passed the 1h-PG threshold for T2D (209 mg/dL, 11.6 mmol/L) first, and 74% passed the 1h-PG threshold for impaired glucose tolerance (IGT; 155 mg/dL) first.

Conclusions:

High 1h-PG NGT is an intermediate risk category between Low 1h-PG NGT and PreDM and may provide an opportunity for early intervention to preserve ß-cell function.

1. Introduction

Early intervention for preventing or delaying type 2 diabetes (T2D) is more effective than treatment (1–3). Hence, early detection of prediabetes (PreDM) and T2D is crucial. Although the diagnosis of PreDM was designed to identify individuals who are at high-risk for developing diabetes, current diagnostic criteria for PreDM are suboptimal (4). About 40 percent of people who progressed to T2D within 10 years had normal glucose tolerance (NGT) at baseline (5) without evidence of PreDM. Therefore, current guidelines using fasting plasma glucose (FPG) and two-hour plasma glucose (2h-PG) during an oral glucose tolerance test (OGTT) to identify individuals with PreDM are suboptimal because a large subset of people who are at high-risk for diabetes but have a normal FPG and 2h-PG may be overlooked (6). The one-hour plasma glucose (1h-PG) is a promising early biomarker for risk stratification to identify people at high-risk earlier in the lengthy trajectory to diabetes while β-cell function is still relatively intact.

The San Antonio Heart Study (SAHS) reported that a 1h-PG threshold of 155 mg/dL (8.6 mmol/L) for impaired glucose tolerance (IGT) predicted risk of T2D in the following 7-8 years with higher accuracy than the standard 2h-PG threshold of 140 mg/dL (7). Superiority or equality of 1h-PG versus 2h-PG in predicting future diabetes was replicated in studies of various ethnicities, including East Asians (4; 6; 8; 9). Furthermore, studies (10–12) also showed that 1h-PG≥155 mg/dL (8.6 mmol/L) predicted cardiovascular disease and mortality. These findings were based on 1h-PG levels measured at baseline, but additional information can be obtained from glycemic trajectories over time and metabolic parameters during follow-up. To investigate the latter, a longitudinal data set derived over a long time with short intervals between follow-up visits is necessary.

Therefore, we conducted a secondary analysis of a 10-year longitudinal study during which OGTTs were carried out every two years. We used the longitudinal data set to 1) validate increased risk of High 1h-PG for T2D; 2) identify longitudinal changes in insulin sensitivity and β-cell function; and 3) show that High 1h-PG is an earlier marker than the current 2h-PG thresholds for IGT and T2D.

2. Participants, materials and methods

2.1. Study population

The current study was based on the Korean Genome Epidemiology Study (KoGES), a population-based epidemiology study designed to assess the incidence and risk factors for non-communicable chronic diseases such as T2D, hypertension, obesity, and cardiovascular disease (CVD). KoGES was composed of two cohorts; the urban Ansan cohort and the rural Ansung cohort. The baseline survey of the KoGES Ansan-Ansung study was completed between 2001 and 2002, with follow-up surveys conducted biennially thereafter for 10 years. Details of the KoGES study and the methods have been previously described (13; 14). Eligibility criteria were age 40-69 years, residence within the borders of the survey area for at least 6 months before enrollment, and mental and physical abilities permitting participation.

The Ansan–Ansung study protocol was approved by the Institutional Review Board of the Korea Centers for Disease Control and Prevention; study participation was voluntary, with written informed consent obtained. The Institutional Review Board of Pusan National University (South Korea) approved the study protocol (IRB No. 2101-007-099).

Among 10,030 participants from KoGES, we identified 8,794 without T2D at baseline (n=1,236 were excluded because they either had diabetes or were taking glucose-lowering medications). Participants who had illnesses such as malignancy or diseases that could affect plasma glucose levels were excluded (n=700). Additionally, participants who had missing glucose or insulin values at t=0, 60, or 120 min during the OGTT at baseline were not eligible for inclusion (n=87). Participants with fewer than two follow-up OGTTs were also excluded (n=2,065). As a result, 6,144 participants were selected (see Supplemental Fig. S1). We selected a subset of 215 participants with NGT and 1h-PG < 209 mg/dL (11.6 mmol/L) and therefore without diabetes at baseline based on 1h-PG criterion, who progressed to IGT and T2D (2h-PG ≥ 200 mg/dL, 11.1 mmol/L) during the study and had at least three OGTTs. Their trajectories were tracked by linear mixed effect model fits, estimating time to T2D and calculating 1h-PG thresholds of IGT and T2D as first introduced in (15).

2.2. Measurement of anthropometric and biochemical parameters

Participants completed a self-administered questionnaire including medical history and sociodemographic information. Biochemical measurements including FPG, fasting insulin, total cholesterol, triglycerides, and high-density lipoprotein cholesterol (HDL-C) were measured enzymatically in plasma using a 747 Chemistry Analyzer (Hitachi, Tokyo, Japan) after a 12-hour overnight fast. Plasma glucose concentrations were measured by the hexokinase method. Plasma insulin concentrations were measured by radioimmunoassay. At the baseline and biennial visits, each participant underwent a 75 g OGTT after a 12-h overnight fast, and plasma glucose and insulin concentrations were collected at t = 0, 60, and 120 min.

2.3. Definitions of NGT, IGT, PreDM, T2D, iIFG, iIGT, CGI, IFG−, IFG+, IGT−, IGT+, CGI−, CGI+, and dysglycemia

Following ADA criteria, NGT was defined by FPG concentration <100 mg/dL (5.6 mmol/L) and 2h-PG concentration <140mg/dL (7.8 mmol/L). IGT was defined by FPG<126 mg/dL (7.0 mmol/L) and 2h-PG ≥140 (7.8 mmol/L) and < 200 mg/dL (11.1 mmol/L). PreDM was defined by: 1) isolated Impaired Fasting Glucose (iIFG): FPG ≥100 mg/dL (5.6 mmol/L) and <126 mg/dL (7.0 mmol/L), and 2h-PG <140mg/dL (7.8 mmol/L); or 2) isolated Impaired Glucose Tolerance (iIGT): FPG < 100 mg/dL (5.6 mmol/L) and 2h-PG ≥140 (7.8 mmol/L) and < 200 mg/dL (11.1 mmol/L); or 3) Combined Glucose Intolerance (CGI): FPG ≥100 (5.6 mmol/L) and < 126 mg/dL (7.0 mmol/L) and 2h-PG ≥140 (7.8 mmol/L) and < 200 mg/dL (11.1 mmol/L). T2D was defined by: 1) FPG ≥ 126 mg/dL (7.0 mmol/L); or 2) 2h-PG ≥ 200mg/dL (11.1 mmol/L); or 3) current treatment with insulin or oral antidiabetic drugs. The three subcategories of PreDM, iIFG, iIGT, and CGI were further subdivided based on 1h-PG glucose. IFG− stands for iIFG with 1h-PG<155 mg/dL (8.6 mmol/L), IFG+ for iIFG with 1h-PG≥155 mg/dL (8.6 mmol/L), IGT− for iIGT with 1h-PG<155 mg/dL (8.6 mmol/L) , IGT+ with 1h-PG≥ 155 mg/dL (8.6 mmol/L), CGI− for CGI with 1h-PG<155 mg/dL (8.6 mmol/L), and CGI+ with 1h-PG> 155 mg/dL (8.6 mmol/L). Dysglycemia was defined as PreDM or T2D, defined above.

2.4. Insulin sensitivity and β-cell function

The Insulin Sensitivity and Secretion (ISS) model was used with Matlab (v.9.5.0 (R2021b), Natick, MA: The MathWorks Inc) to estimate insulin sensitivity mSI and β-cell function (mBCF) at each follow-up visit (16). In addition, β-cell function was assessed using the homeostasis model assessment of β-cell function (HOMA-β) and 60-min insulinogenic index (IGI60) (17). HOMA-β was calculated as 360 x fasting insulin (μU/ml)/ (fasting glucose (mg/dL) – 63), and IGI60 was calculated as (plasma insulin at 60 min – plasma insulin at 0 min)/(plasma glucose at 60 min – plasma glucose at 0 min). HOMA insulin resistance (HOMA-IR) was calculated as 405/(fasting insulin (μU/ml) x fasting glucose (mg/dL) (18). The Matsuda index was adopted as an additional measure of insulin sensitivity and calculated as 10000/sqrt(FPG x fasting plasma insulin x mean plasma glucose x mean plasma insulin) using three time points during the OGTT. The oral disposition index (oDI) was calculated as Matsuda index x IGI60 (19).

2.5. Statistical analysis

Categorical variables are presented as frequency and percentages. Continuous variables with normal distribution are presented as mean and standard deviation (SD); skewed variables are reported as median and interquartile range (IQR). Comparisons of two groups were done by two-sample t-test or Wilcoxon rank-sum test for continuous data and chi-squared test for categorical data. Comparisons among three groups were done by one-way ANOVA or the Kruskal-Wallis test for continuous data and the chi-squared test for categorical data. The post hoc test was done by the Bonferroni method. Cumulative incidence, as 1 – survival, was computed using the Kaplan-Meier method. A Cox proportional hazards model was used to estimate hazard ratios (HR) and 95% confidence intervals (CI). Survival analyses were conducted using the R packages survival and survminer.

We used a linear mixed effect model (LME) to correlate 1h-PG with 2h-PG, 1h-PG with time (year) and 2h-PG with time (year) as done previously (15). The individual fits were used to determine times when these measures crossed their respective thresholds. LMEs were done with the R Statistical language (version 4.0.4; R Core Team, 2021), using function lmer in the lme4 library. R² for the fits was evaluated using the MuMIn package, and p values were determined using the R function anova to compare models with and without the fixed effects. The statistical model for LME in Wilkinson notation was y ~ x + (x | participant), where y is either 1h-PG or 2h-PG, and x is either 2h-PG or the individual’s follow-up time. That is, the fixed effects were slope and intercept, and each individual had their own random intercept and slope. A P-value < 0.05 was considered statistically significant without Bonferroni correction. When Bonferroni correction was applied to adjust for multiple comparisons, the threshold for statistical significance was adjusted to 0.05/the number of comparisons.

3. Results

3.1. Baseline Characteristics

We identified 6,144 participants without diabetes at baseline (Table 1). There were 3,654 (59.3%) classified as Low 1h-PG, 1,190 (19.2%) as High 1h-PG, and 1,309 (21.3%) as PreDM. 5.4% (197/3645) of Low 1h-PG, 20% (238/1190) of Low 1h-PG, and 43.6% (571/1309) of PreDM progressed to T2D during the study. Body mass index (BMI), total cholesterol, LDL-cholesterol, and triglycerides increased from Low 1h-PG to High 1h-PG to PreDM. Time to follow-up, the number of OGTTs, and OGTT intervals were not significantly different across the three groups (Table 1). A subset of 215 participants who progressed from NGT to IGT to T2D and had 1h-PG<209 mg/dL (11.6 mmol/L) at baseline showed no significant differences in the numbers of visits, OGTTs, and follow-up durations between Low 1h-PG and High 1h-PG (Supplemental Table S1).

Table 1.

Anthropometric, clinical, and metabolic characteristics of study participants at baseline (N=6,144)

Characteristics	Overall (N=6,144)	Low 1h-PG (N = 3,645)	High 1h-PG (N = 1,190)	Prediabetes (N=1,309)	P value
Age, years	51.3 (8.6)	50.6 (8.4)	52.5 (8.7)†	52.3 (8.9)†	<0.001
Sex (female), n (%)	3039 (49.5)	1688 (46.3)	707 (59.4)†	644 (49.2)‡	<0.001
BMI, kg/m2	24.5 (3.1)	24.2 (2.9)	24.6 (3.2)†	25.0 (3.2)†‡	<0.001
Waist circumference, cm	82.2 (8.6)	81.3 (8.3)	83.3 (8.9)†	83.5 (8.9)†	<0.001
SBP	116 (17)	114 (16)	118 (17)†	118 (18)†	<0.001
DBP	74 (11)	73 (11)	76 (11)†	75 (11)†	<0.001
Laboratory findings
Creatinine, mg/dL	0.84 (0.19)	0.83 (0.19)	0.86 (0.17)†	0.9 (0.2)†	<0.001
Total cholesterol, mg/dL	190.0 (33.9)	185.9 (33.0)	192.7 (34.1)†	198.85 (34.4)†‡	<0.001
LDL, mg/dL	116.1 (32.1)	113.9 (31.2)	117.3 (34.0)†	120.97 (32.2)†‡	<0.001
HDL, mg/dL	44.9 (9.9)	44.7 (9.6)	45.6 (10.3)†	45.0 (10.3)	0.06
Triglyceride, mg/dL	145.0 (98.9)	136.7 (93.1)	149.3 (105.2)†	164.3 (105.3)†‡	<0.001
OGTT measures
Glucose (0 min), mg/dl	83 (8)	80 (6)	84 (7)†	88 (11)†‡	<0.001
Glucose (60 min), mg/dl	141 (40)	116 (23)	178 (20)†	177 (35)†	<0.001
Glucose (120 min), mg/dl	113 (29)	100 (19)	110 (21) †	154 (22)†‡	<0.001
Insulin (0 min), μU/ml	7.5 (4.4)	7.4 (4.4)	7.4 (5.3)	7.7 (3.7)†	<0.01
Insulin (60 min), μU/ml	32.8 (31.9)	29.0 (26.9)	44.3 (41.7)†	32.9 (31.9)†‡	<0.001
Insulin (120 min), μU/ml	28.1 (27.1)	23.2 (19.6)	30.0 (28.7)†	40.3 (37.6)†‡	<0.001
Follow-up duration, year	9.3 (1.6)	9.3 (1.6)	9.3 (1.5)	9.3 (1.6)	0.86
Number of follow-ups	5.4 (0.9)	5.4 (0.9)	5.5 (0.9)	5.4 (0.9)	0.674
OGTT interval, year	2.3 (0.5)	2.3 (0.5)	2.3 (0.6)	2.3 (0.5)	0.331
Metabolic estimates
HOMA-IR	1.5 (0.9)	1.4 (0.8)	1.5 (1.1)†	1.7 (0.9)†‡	<0.001
HOMA-β	164.0 (155.0)	184.0 (173.2)	140.0 (131.1)†	130.1 (104.4)†‡	<0.001
Matsuda index	11.7 (9.8)	13.0 (10.0)	10.1 (7.4)†	9.5 (10.5)†	<0.001
IGI	0.8 (2.4)	1.1 (3.1)	0.5 (0.4)†	0.4 (0.5)†‡	<0.001
mSI	1.2 (0.6)	1.3 (0.4)	1.1 (0.7)†	1.1 (0.7)†	<0.001
mBCF	0.3 (0.2)	0.41 (0.25)	0.2 (0.1)†	0.1 (0.1)†‡	<0.001

Variables are presented as number (%) or mean (SD). BMI, body mass index; SBP, systolic blood pressure; DBP, diastolic blood pressure; HbA1c, glycated hemoglobin; LDL, low density lipoprotein; HDL, high density lipoprotein; HOMA-IR, homeostatic Model Assessment for Insulin Resistance; HOMA-β, homeostatic Model Assessment for β cell function; IGI, insulinogenic index. OGTT, Oral glucose tolerance test. mSI, model-derived sensitivity; mBCF, model-derived beta-cell function.

^†P<0.016 vs. Low 1h-PG.

^‡P<0.016 vs. High 1h-PG, after Bonferroni correction.

3.2. Survival Analyses

Overall, 16.3 % (1006/ 6,144) progressed to T2D over 10 years of follow-up with highest rates of progression in individuals who had PreDM at baseline. Of 1,190 participants in the High 1h-PG group, 238 (20.0%) progressed to T2D, while 197 of 3,645 (5.4%) in the Low 1h-PG group and 571 of 1,309 (43.6%) in the PreDM group progressed to T2D. Kaplan-Meier survival analysis showed that the High 1h-PG group had a higher risk of developing T2D (HR 3.9, 95% CI 3.2–4.7, P<0.001) compared to the Low 1h-PG group (Fig. 1), but a lower risk of developing T2D than the PreDM group (HR 10.0, 95% CI 8.5-11.7, P<0.001).

Fig. 1.

Survival analysis to T2D: Low 1h-PG vs. High 1h-PG vs. PreDM. The hazard ratios (HR) determined by the Kaplan-Meier curves for progression to T2D of High 1h-PG and PreDM vs. Low 1h-PG are 3.9 [3.3, 4.7] and 10.0 [8.5, 11.7].

We also assessed the risk of T2D defined by the newly proposed IDF criterion for post-load glucose, 1h-PG ≥ 209 mg/dL (11.6 mmol/L). For this analysis, participants without diabetes at baseline who had 1h-PG ≥ 209 mg/dL (11.6 mmol/L) were excluded (n=334). With an outcome of T2D defined as FPG≥126 mg/dL (7.0 mmol/L) or 1h-PG≥209 mg/dL (11.6 mmol/L), Kaplan-Meier survival analysis showed that the High 1h-PG group had a higher risk of developing T2D (HR 4.3, 95% CI 3.8–4.9, P<0.001) compared to the Low 1h-PG group, but a comparable risk of developing T2D to the PreDM group (HR 4.8, 95% CI 4.2-5.4, P<0.001) (Supplemental Fig. S2).

For further survival analysis of T2D risk, we subdivided PreDM into iIGT, iIFG, and CGI (Supplemental Fig. S3A). Compared to Low 1h-PG, CGI had the highest risk of developing T2D (HR 23.1, 95% CI 18-29.7, P<0.001), followed by iIGT (HR 9.1, 95% CI 7.7-10.7, P<0.001), iIFG (HR 8.7, 95% CI 6.5-11.7, P<0.001), and High 1h-PG (HR 3.9, 95% CI 3.2-4.7, P<0.001). To further determine the effect of High 1h-PG on risk for T2D, the above subtypes of PreDM (iIGT, iIFG, or CGI) were subdivided into participants without and with High 1h-PG: IGT− vs. IGT+, IFG− vs. IFG+, and CGI− vs. CGI+, respectively (Supplemental Fig. S3B). Each subtype of High 1h-PG group had a higher risk for T2D than the corresponding participants with Low 1h-PG (HR 10.3 [8.6, 12.3] vs. 6.0 [4.6, 7.8], 10.7 [7.8, 14.8] vs. 4.7 [2.5, 8.9], and 24.5 [19.0, 31.5] vs 6.8 [1.7, 27.5], P<0.001 for all).

3.3. Longitudinal changes of metabolic parameters

Longitudinal changes in metabolic parameters of model-derived insulin sensitivity (mSI) and β-cell function (mBCF) (n=6144 at baseline) in the Low 1h-PG, High 1h-PG and PreDM groups are shown in Fig. 2. Insulin sensitivity decreased over time in all three groups, but mSI was higher in the Low 1h-PG than in the High 1h-PG and PreDM groups at baseline and follow-up. mSI in PreDM was comparable to High 1h-PG at baseline but exhibited more marked longitudinal decrease than High 1h-PG (Fig. 2A; see Supplemental Table S2 for details of statistics). Low 1h-PG had the strongest mBCF, PreDM the weakest, and High 1h-PG was intermediate at baseline and during follow-up (Fig. 2B; see Supplemental Table S2). mBCF started low in High 1h-PG and PreDM and stayed low, while it started higher and stayed higher in Low 1h-PG even though it declined with time. The patterns of the Matsuda and IGI60 indices were similar to mSI and mBCF (Supplemental Fig. S4 and Supplemental Table S3). The metabolic parameters mSI and mBCF in the subtypes of NGT and preDM groups further subdivided by High 1h-PG are shown in Supplemental Fig. S5. At baseline and follow-up, model-derived insulin sensitivity (mSI) was higher in Low 1h-PG than High 1h-PG, but there were no significant differences between IGT− and IGT+, IFG− and IFG+, and CGI− and CGI+, respectively (Supplemental Fig. S5A, Supplemental Table S4). β-cell function (mBCF) was higher in each Low 1h-PG group than the corresponding High 1h-PG group (Supplemental Fig. S5B, Supplemental Table S4). The patterns of the Matsuda and IGI 60 indices were similar to mSI and mBCF (not shown).

Fig. 2.

Longitudinal changes of metabolic parameters in Low 1h-PG, High 1h-PG, and PreDM. (A) The longitudinal changes of model-derived insulin sensitivity (mSI) (16). mSI is higher in Low 1h-PG than High 1h-PG and PreDM at baseline and all follow-up times. PreDM has comparable SI at baseline as High 1h-PG but exhibits worsening longitudinal decline, compared to High 1h-PG. (B) The longitudinal changes of model-derived β-cell function (mBCF) (16). mBCF is the highest in Low 1h-PG, intermediate in High 1h-PG, and the lowest in PreDM at baseline and all follow-up times. See Supplemental Table S2 for detailed statistics.

3.4. Times to IGT and T2D from Low 1h-PG and High 1h-PG

In the subset of 215 participants who progressed from NGT to IGT and T2D and had 1h-PG<209 mg/dL (11.6 mmol/L) at baseline, times to onset of IGT and T2D in Low and High 1h-PG groups were estimated by LME, using the methods introduced in (15).Individual LME fits of 2h-PG vs. time (dotted lines in Supplemental Fig. S6A) were used to estimate times of crossing the IGT (140 mg/dL) (7.8 mmol/L) and T2D (200 mg/dL) (11.1 mmol/L) thresholds. The times were calculated from the times the fitted lines intercepted the two horizontal lines in each box. This information was used to compare the times to progression for those who were in the High 1h-PG and Low 1h-PG groups at baseline. High 1h-PG progressed to IGT 0.2 years earlier (2.0 (0.5) vs. 1.8 (0.5) years, P=0.01) and to T2D 0.9 years earlier (9.1 (2.5) vs. 8.2 (2.6) years, P=0.01), mean (SD)) than Low 1h-PG, as summarized in Fig. 3.

Fig. 3.

Estimated times from Low 1h-PG and High 1h-PG to reach IGT and T2D. A) Times to IGT from Low-1h PG and High 1h-PG are 2.0 (0.5) years and 1.8 (0.5) years, respectively, resulting in a 0.2 year time gap, P=0.01. B) Times to T2D from Low-1h PG and High 1h-PG are 9.1 (2.6) years and 8.2 (2.6) years, respectively, resulting in a 0.9 year time gap, P=0.01, mean (SD).

3.5. Times to the respective 1h-PG and 2h-PG thresholds of IGT and T2D

In addition to times to the 2h-PG thresholds of IGT and T2D described above, individual LME fits of 1h-PG vs. time (dotted lines of Supplemental Fig. S6B) were used to estimate the times to cross the proposed 1h-PG thresholds for IGT and T2D, 155 mg/dL (8.6 mmol/L) and 209 mg/dL (11.6 mmol/L), respectively. Participants reached the 1h-PG threshold of 155 mg/dL (8.6 mmol/L) 2.6 (1.7) years earlier than the 2h-PG threshold 140 mg/dL (7.8 mmol/L) of IGT, and the 1h-PG threshold 209 mg/dL (11.6 mmol/L) 1.4 (3.8) years, mean (SD) earlier than the 2h-PG threshold 200 mg/dL (11.1 mmol/L) of T2D. To illustrate the order of crossing their respective 1h-PG and 2h-PG thresholds, the pairs of crossing times were plotted against each other for each participant in Supplemental Fig. S7. 92% (197/215) crossed the 1h-PG threshold for IGT first (Supplemental Fig. S7A), and 74% (160/215) crossed the 1h-PG threshold for T2D first (Supplemental Fig. S7B).

3.6. Order of crossing 1h-PG and 2h-PG thresholds for IGT and T2D

A linear mixed effect model was also used to estimate the correlation between 1h-PG and 2h-PG (Supplemental Fig. S6C; R = 0.76, P <0.001). Each individual’s LME fitted line of 1h-PG vs. 2h-PG in Supplemental Fig. S6C was used to estimate the 1h-PG values when the thresholds for IGT and T2D based on 2h-PG were reached. The 1h-PG values were determined by intersection of the fitted line with the two horizontal lines, 2h-PG = 140 mg/dL (7.8 mmol/L) and 2h-PG =200 mg/dL (11.1 mmol/L), respectively. The estimated 1h-PG values for IGT and T2D were 162 mg/dL (9.0mmol/L) and 244 mg/dL (13.5mmol/L), respectively (Supplemental Fig. S8A and B). Likewise, the 2h-PG values when the fitted line meets the two vertical lines, 1h-PG = 155 mg/dL (8.6 mmol/L) and 1h-PG = 209 mg/dL (11.6 mmol/L) were calculated. To further illustrate the order of crossing the prospective thresholds, the pairs of estimated 1h-PG and 2h-PG values are plotted in Supplemental Fig. S8C and D. Participants in the fourth quadrant passed the 1h-PG thresholds first for IGT (159 out of 215, 74%) and T2D (214 out of 215, 99%).

4. Discussion

Several studies have shown that people with 1h-PG ≥ 155 mg/dL (8.6 mmol/L) are at high risk for future T2D (4; 7; 20). However, these studies did not use frequently collected longitudinal OGTTs to confirm high-risk status. Here we used the 10-year longitudinal data set of biannually collected OGTTs from a cohort of Koreans during the KoGES study to confirm that NGT with High 1h-PG (≥155 mg/dL [8.6 mmol/L]) at baseline is an intermediary hyperglycemic state representing a higher risk for future T2D than NGT with Low 1h-PG <155 mg/dL (8.6 mmol/L) (HR 3.9, 95% CI 3.3–4.7, P<0.001) but lower than PreDM at baseline (HR 10.0, 95% CI 8.5–11.7, P<0.001). The greater risk of people with NGT and High 1h-PG vs people with NGT and Low 1h-PG was reflected in the times to reach IGT (0.2 years shorter) and T2D (0.9 years shorter) (Fig. 3).

4.1. Metabolic characteristics and clinical implications

Insulin sensitivity and β-cell function were lower in NGT with High 1h-PG than NGT with low 1h-PG at every time point although insulin sensitivity decreased in the three groups over time. Since individuals with High 1h-PG are in an intermediary state with relatively preserved β-cell function compared to individuals with PreDM, early interventions to optimize insulin sensitivity could delay or prevent progression to T2D. The current findings in KoGES study are consistent with our previous longitudinal findings in Southwest Native Americans (SWNA) (15) and previous cross-sectional studies (4; 7; 20) in which High 1h-PG was more associated with β-cell function than insulin sensitivity within those with PreDM.

4.2. Order of crossing 1h-PG vs. 2h-PG

Our findings that the majority of participants passed the 1h-PG threshold first for IGT (159 out of 215, 74%) and almost all participants passed 1h-PG threshold first for T2D (214 out of 215, 99%) are consistent with our previous findings in SWNA (15). In both groups, the thresholds of 1h-PG for IGT and T2D were crossed earlier than those of 2h-PG, and High 1h-PG with NGT was an intermediate state between Low 1h-PG with NGT and IGT.

4.3. Diabetes conversion rate

By taking advantage of frequently collected OGTTs, we also confirmed previous findings in the literature that about 40% of progressors to T2D over a 10-year follow-up had NGT at baseline (2). The current study similarly showed that about 43.2% of people diagnosed with T2D during a 10-year longitudinal study had NGT at baseline. Furthermore, 43.6% of PreDM, 20% of High 1h-PG NGT, and 5.4% of Low1h-PG NGT at baseline developed T2D during the 10-year study.

4.4. Strengths and Limitations

As is customary in epidemiologic studies of T2D, a single OGTT was used for diagnosis and modest reproducibility of glucose biomarkers could result in false positive or false negative diagnosis of T2D. Nevertheless, the long follow-up and rigorous 2-year protocol is a study strength and provides robust survival data for estimating risk and time to progression of T2D in this well characterized cohort. The current study had good retention of participants from baseline to year 10 (6144 vs. 4117 [67%]), even though there were losses to follow-up partly due to progression to T2D. HbA_1c was not collected at early follow-ups because the study was launched in 2002 when HbA_1c had not been standardized as a criterion of dysglycemia. As the Matsuda index was calculated based on three glucose and insulin data points (t = 0, 60, 120 min), it may not represent insulin sensitivity as well as the standard set of five data points (t=0, 30, 60, 90, 120 min).

4.5. Conclusion

People with NGT and High 1h-PG (≥155 mg/dL) (8.6 mmol/L) are at intermediate risk for T2D between NGT with Low 1h-PG (<155 mg/dL) (8.6 mmol/L) and PreDM. Moreover, the majority (74%) of progressors from NGT with Low 1h-PG to T2D passed through the intermediate state of NGT with High 1h-PG≥155 mg/dL (8.6 mmol/L), characterized by β-cell dysfunction and insulin resistance, compared to people with NGT and Low 1h-PG. NGT with High 1h-PG is characterized by better β-cell function than PreDM but comparable insulin resistance at baseline. Screening with 1h-PG during an OGTT can identify high-risk people earlier in the time course to T2D who may benefit from lifestyle intervention when the β-cell is more functional.