In diabetes, low concentrations of the biomarker 1,5-anhydroglucitol (1,5-AG) reflect hyperglycemic excursions over the prior 1–2 weeks. To the extent that hyperglycemic excursions are important in atherogenesis, 1,5-AG may provide independent information regarding cardiovascular risk. Nonetheless, few studies have evaluated associations of 1,5-AG with long-term cardiovascular outcomes in a population-based setting. We measured 1,5-AG in 11,106 participants in the Atherosclerosis Risk in Communities (ARIC) study without cardiovascular disease at baseline (1990–1992) and examined prospective associations with coronary heart disease (n = 1,159 events), ischemic stroke (n = 637), heart failure (n = 1,553), and death (n = 3,120) over 20 years of follow-up. Cox proportional hazards models were adjusted for demographic and cardiovascular risk factors. Compared with persons with 1,5-AG ≥6 μg/mL and no history of diabetes, persons with diabetes and 1,5-AG <6.0 μg/mL had an increased risk of coronary heart disease (HR 3.85, 95% CI 3.11–4.78), stroke (HR 3.48, 95% CI 2.66–4.55), heart failure (HR 3.50, 95% CI 2.93–4.17), and death (HR 2.44, 95% CI 2.11–2.83). There was a threshold effect, with little evidence for associations at “nondiabetic” concentrations of 1,5-AG (e.g., >10 μg/mL). Associations remained but were attenuated with additional adjustment for fasting glucose or HbA1c. These data add to the growing evidence for the prognostic value of 1,5-AG for long-term complications in the setting of diabetes.

Hemoglobin A1c (HbA1c) reflects glycemic exposure over the past 2–3 months, is the standard measure used for the clinical monitoring of glucose control, and is also recommended for diagnosis of diabetes (1). 1,5-Anhydroglucitol (1,5-AG) or 1-deoxyglucose is a monosaccharide originating primarily from dietary sources and is an alternative biomarker of hyperglycemia. In the normal state, 1,5-AG is typically present at high but constant concentrations in the blood. It is freely filtered by the glomeruli and reabsorbed in the renal tubule with a small amount, corresponding to dietary intake, excreted in the urine. In the setting of hyperglycemia (specifically, when blood glucose exceeds the renal threshold of ∼160–180 mg/dL), high amounts of glucose block renal tubular reabsorption of 1,5-AG, causing serum 1,5-AG concentrations to fall. Therefore, low serum 1,5-AG can serve as a marker of short-term hyperglycemia, and concentrations are thought to reflect hyperglycemic episodes over a period of ∼1–2 weeks (24). Appealingly, 1,5-AG is a nonfasting test that may capture additional information on glycemic excursions that are not reflected in HbA1c (5). A growing literature provides evidence that 1,5-AG may provide a useful complement to HbA1c measurements in some settings (514), especially when one seeks to characterize short-term glycemic variability that may not be reflected in standard metrics of glycemia.

Previous studies suggest that postprandial glycemic excursions may be an independent risk factor for cardiovascular disease (1521), although this contention is controversial (2225). Chronic exposure to postprandial elevations in glucose is hypothesized to induce endothelial dysfunction and contribute to the development of atherosclerosis (26,27). There is some evidence from epidemiologic studies that 2-h glucose measurements may be more strongly associated with cardiovascular events compared with fasting glucose (24,28), but this finding has been inconsistent (29). Among persons without a history of diabetes, HbA1c is more strongly associated with vascular outcomes compared with fasting glucose (3032). This may partly reflect the lower within-person variability of HbA1c compared with fasting glucose but may also be a function of the importance of nonfasting glucose in the development of vascular complications of diabetes (33).

To the extent that hyperglycemic excursions are important in atherogenesis, 1,5-AG may provide independent information regarding cardiovascular risk. Nonetheless, few studies have evaluated the associations of 1,5-AG with long-term cardiovascular outcomes in a population-based setting. We have previously shown that 1,5-AG is strongly associated with important microvascular outcomes (kidney disease and retinopathy), particularly in persons with diabetes and even after adjustment for HbA1c (34). The association of 1,5-AG with incident cardiovascular outcomes is uncharacterized. In this study, our aim was to characterize the independent association of 1,5-AG with future risk for coronary heart disease, heart failure, stroke, and all-cause mortality in a community-based population.

Study Population

The Atherosclerosis Risk in Communities (ARIC) study is a community-based prospective cohort of >15,000 participants sampled from four U.S. communities. The first clinic examinations (visit 1) took place from 1987 to 1989, with three follow-up visits approximately every 3 years (35). A fifth visit was recently completed (2011–2013). The second clinic examination (visit 2) took place from 1990 to 1992 and is the baseline for the current study. There were 14,348 participants who attended visit 2. Institutional review boards at all institutions reviewed the study, and informed consent was obtained from all participants.

In the current study, we excluded all persons whose race/ethnicity was recorded as other than white or black (N = 91); who were fasting <8 h (N = 446); who had a history of coronary heart disease, stroke, or heart failure (N = 1,391); or who were missing variables of interest (N = 1,314) for a final analytic sample of 11,106 (762 persons with diagnosed diabetes and 10,344 persons without a diagnosis of diabetes). Persons with diagnosed diabetes were classified on the basis of a self-reported history of physician-diagnosed diabetes or current glucose-lowering medication use.

Measurement of 1,5-AG

1,5-AG (GlycoMark, Winston-Salem, NC) was measured using a Roche Modular P800 system in 2012–2013 in stored serum specimens obtained at ARIC visit 2. The interassay CV was 5%. The reliability coefficient for N = 610 masked duplicate specimen pairs was 0.99. Previous studies have shown this 1,5-AG assay to be highly reliable even in long-term stored samples (8,36).

Other Variables

Serum glucose was measured using the hexokinase method. HbA1c was measured in whole blood samples using high-performance liquid chromatography with instruments standardized to the Diabetes Control and Complications Trial assay (Tosoh A1c 2.2. Plus Glycohemoglobin and Tosoh G7 analyzers) (37). Plasma lipid concentrations, BMI (measured as weight in kilograms divided by the square of height in meters), and blood pressure were measured using standard ARIC protocols (3842). Serum creatinine was measured using a modified kinetic Jaffé method. Estimated glomerular filtration rate was calculated from serum creatinine using the 2009 CKD-Epidemiology Collaboration (CKD-EPI) equation (43). Hypertension was defined as the mean of the second and third readings at the visit (with cutoff for systolic blood pressure of 140 mmHg or higher and/or a cutoff for diastolic blood pressure of 90 mmHg or higher) or the use of hypertension medication. Education, alcohol use, and smoking status were self-reported. Physical activity was assessed using the Baecke index, a measure of habitual leisure (sport- and exercise-related) activity (44).

Assessment of Coronary Heart Disease, Stroke, Heart Failure, and All-Cause Mortality

The ascertainment of deaths and classification of cardiovascular events are detailed elsewhere (45,46). Briefly, deaths and potential cardiovascular hospitalizations were reported annually by participants (or proxy) and also identified through community-wide hospital surveillance and linkage to state and national death indexes. Trained personnel abstracted hospital records related to possible cardiovascular events, and these outcomes are adjudicated by a panel of experts. Silent myocardial infarction, as detected by means of electrocardiography during the visits, was also identified and recorded. We defined newly diagnosed coronary heart disease as a definite or probable myocardial infarction, a death from coronary heart disease, or electrocardiographic evidence of a silent myocardial infarction detected at one of the follow-up visits. We also examined definite or probable ischemic stroke (adjudicated). Incident heart failure was defined as the first heart failure hospitalization identified by ICD-9 codes of 428.X in any position on the hospital discharge list or a death certificate with death from heart failure in any position. Follow-up data for all cardiovascular events were available up to 1 January 2013.

Statistical Analyses

Baseline characteristics of the study population were compared across categories of 1,5-AG in persons with and without a history of diagnosed diabetes. Low serum concentrations of 1,5-AG reflect hyperglycemic excursions (inverse association with serum glucose); a 1,5-AG concentration of ≤6 μg/mL is thought to reflect high peaks of glucose (higher than ∼200 mg/dL) over the past 1–2 weeks, whereas a concentration of ≥10 is thought to reflect the absence of recent significant hyperglycemia (glucose peaks lower than ∼180 mg/dL) (5). We divided the populations of persons with and without diabetes into two groups based on a cut point of 6 μg/mL. Those with no history of diabetes and 1,5-AG ≥6 μg/mL served as the common reference group in our overall categorical analysis. We also conducted analyses stratified by diabetes status.

To characterize the associations of 1,5-AG with incident cardiovascular outcomes and all-cause mortality, we used Cox proportional hazards models to estimate hazard ratios (HRs) and their corresponding 95% CIs. We verified that the proportional hazards assumption was met using log-log plots. In analyses with 1,5-AG modeled categorically, P values for trends were calculated by modeling the category medians as a continuous variable. To characterize the shape of the continuous association of 1,5-AG at baseline with each end point, we fit linear and restricted cubic splines, using the 10th percentile (1,5-AG = 10 μg/mL) as the reference point and with knots placed at the 5th, 35th, 65th, and 95th percentiles (47).

We constructed four models for each of the outcomes. Model 1 was adjusted for age, sex, race-field center (white participants, MN, MD, and NC; black participants, MS and NC). Model 2 was adjusted for all variables in model 1 plus LDL cholesterol (mg/dL), HDL cholesterol (mg/dL), triglycerides (mg/dL), BMI (kg/m2), waist-to-hip ratio, systolic blood pressure (mmHg), blood pressure–lowering medication use (yes, no), education (less than high school, high school or equivalent, more than high school), drinking status (current, former, never), smoking status (current, former, never), physical activity index, and glomerular filtration rate (mL/min per 1.73 m2). Model 3 was adjusted for all variables in model 2 plus HbA1c (%). Model 4 was adjusted for all variables in model 2 plus fasting glucose (mg/dL). We tested for multiplicative interactions by age, sex, and race. All statistical analyses were conducting using Stata SE, version 13.1.

In persons with diagnosed diabetes (n = 762), almost half (49%) had low 1,5-AG (<6 μg/mL)—consistent with recent peaks of glucose of higher than ∼200 mg/dL; 62% had 1,5-AG <10 μg/mL. In persons without a history of diagnosed diabetes (n = 10,344), just less than 2% (n = 203) of participants had concentrations <6 μg/mL. Categories of 1,5-AG were strongly and inversely associated with traditional diabetes risk factors (Table 1). Among persons with a diagnosis of diabetes, those in the low 1,5-AG category (<6 μg/mL) had higher mean HbA1c and fasting glucose values, were more likely to be obese or have hypertension, and had a poorer lipid profile.

Table 1

Characteristics of participants without a history of cardiovascular disease overall and by categories of 1,5-AG and diagnosed diabetes status at baseline: the ARIC study (N = 11,106)

OverallNo diagnosed diabetes, N = 10,344
Diagnosed diabetes, N = 762
1,5-AG 
≥6 μg/mL1,5-AG 
<6 μg/mL1,5-AG 
≥6 μg/mL1,5-AG 
<6 μg/mL
N 11,106 10,141 203 389 373 
1,5-AG, μg/mL 18.2 (14.3, 21.9) 18.7 (15.2, 22.3) 4.2 (2.2, 5.2) 16 (13.1, 20.1) 1.8 (1.2, 3.2) 
1,5-AG, μg/mL, range 0.6, 49.4 6.0, 49.4 0.6, 5.9 6.0, 33.9 0.6, 5.9 
HbA1c, % 5.4 (5.2, 5.8) 5.4 (5.2, 5.7) 6.3 (5.4, 8.3) 6.0 (5.6, 6.7) 9.6 (8.3, 11.0) 
HbA1c, mmol/mol 36 (33, 40) 36 (33, 39) 45 (36, 67) 42 (38, 50) 81 (67, 97) 
Fasting glucose, mg/dL 102 (96, 111) 101 (95, 109) 126 (100, 205) 124 (105, 147) 235 (192, 294) 
Age, years 57 (52, 62) 57 (52, 62) 58 (53, 64) 58 (54, 64) 58 (53, 63) 
Female, % 58.6 58.5 59.6 58.1 62.7 
Black, % 23.4 21.9 32.5 35.7 47.7 
BMI, kg/m2 27.1 (24.2, 30.6) 27 (24, 30) 28 (25, 33) 29.3 (26.4, 33.6) 31.1 (27.4, 34.4) 
BMI ≥30 kg/m2, % 28.5 26.4 42.9 45.8 57.6 
Hypertension, % 33.3 31.4 43.1 54.4 57.4 
Education, %      
 Less than high school 19.8 18.6 24.6 37.8 30.6 
 High school or equivalent 41.9 42.3 38.4 33.4 42.6 
 College or above 38.3 39.1 37.0 28.8 26.8 
Current alcohol use, % 57.9 59.6 57.6 41.7 30.8 
Current smoking status, % 21.4 21.6 19.7 21.9 17.4 
Physical activity index 2.25 (1.75, 3.00) 2.25 (1.75, 3.00) 2.25 (1.75, 2.75) 2.25 (1.75, 2.75) 2.00 (1.75, 2.75) 
LDL cholesterol, mg/dL 131 (109, 155) 131 (109, 155) 134 (112, 166) 132 (106, 152) 137 (110, 162) 
HDL cholesterol, mg/dL 48 (39, 60) 48 (39, 61) 43 (36, 57) 44 (35, 54) 43 (35, 52) 
Triglycerides, mg/dL 111 (81, 157) 109 (80, 153) 128 (94, 182) 127 (91, 179) 149 (106, 212) 
eGFR, mL/min per 1.73 m2 98 (89, 106) 98 (89, 105) 100 (90, 110) 100 (88, 109) 102 (91, 113) 
eGFR <60 mL/min per 1.73 m2, % 1.37 1.18 3.45 2.83 3.75 
OverallNo diagnosed diabetes, N = 10,344
Diagnosed diabetes, N = 762
1,5-AG 
≥6 μg/mL1,5-AG 
<6 μg/mL1,5-AG 
≥6 μg/mL1,5-AG 
<6 μg/mL
N 11,106 10,141 203 389 373 
1,5-AG, μg/mL 18.2 (14.3, 21.9) 18.7 (15.2, 22.3) 4.2 (2.2, 5.2) 16 (13.1, 20.1) 1.8 (1.2, 3.2) 
1,5-AG, μg/mL, range 0.6, 49.4 6.0, 49.4 0.6, 5.9 6.0, 33.9 0.6, 5.9 
HbA1c, % 5.4 (5.2, 5.8) 5.4 (5.2, 5.7) 6.3 (5.4, 8.3) 6.0 (5.6, 6.7) 9.6 (8.3, 11.0) 
HbA1c, mmol/mol 36 (33, 40) 36 (33, 39) 45 (36, 67) 42 (38, 50) 81 (67, 97) 
Fasting glucose, mg/dL 102 (96, 111) 101 (95, 109) 126 (100, 205) 124 (105, 147) 235 (192, 294) 
Age, years 57 (52, 62) 57 (52, 62) 58 (53, 64) 58 (54, 64) 58 (53, 63) 
Female, % 58.6 58.5 59.6 58.1 62.7 
Black, % 23.4 21.9 32.5 35.7 47.7 
BMI, kg/m2 27.1 (24.2, 30.6) 27 (24, 30) 28 (25, 33) 29.3 (26.4, 33.6) 31.1 (27.4, 34.4) 
BMI ≥30 kg/m2, % 28.5 26.4 42.9 45.8 57.6 
Hypertension, % 33.3 31.4 43.1 54.4 57.4 
Education, %      
 Less than high school 19.8 18.6 24.6 37.8 30.6 
 High school or equivalent 41.9 42.3 38.4 33.4 42.6 
 College or above 38.3 39.1 37.0 28.8 26.8 
Current alcohol use, % 57.9 59.6 57.6 41.7 30.8 
Current smoking status, % 21.4 21.6 19.7 21.9 17.4 
Physical activity index 2.25 (1.75, 3.00) 2.25 (1.75, 3.00) 2.25 (1.75, 2.75) 2.25 (1.75, 2.75) 2.00 (1.75, 2.75) 
LDL cholesterol, mg/dL 131 (109, 155) 131 (109, 155) 134 (112, 166) 132 (106, 152) 137 (110, 162) 
HDL cholesterol, mg/dL 48 (39, 60) 48 (39, 61) 43 (36, 57) 44 (35, 54) 43 (35, 52) 
Triglycerides, mg/dL 111 (81, 157) 109 (80, 153) 128 (94, 182) 127 (91, 179) 149 (106, 212) 
eGFR, mL/min per 1.73 m2 98 (89, 106) 98 (89, 105) 100 (90, 110) 100 (88, 109) 102 (91, 113) 
eGFR <60 mL/min per 1.73 m2, % 1.37 1.18 3.45 2.83 3.75 

Continuous variables are median (25th, 75th percentiles) unless otherwise noted. eGFR, estimated glomerular filtration rate.

The substantially different distributions of 1,5-AG in persons with and without a diagnosis of diabetes are shown in the histograms (Fig. 1). In persons without diagnosed diabetes (the majority of participants in this study), the distribution of 1,5-AG is roughly normal (light gray histogram). In persons with diagnosed diabetes, the distribution of 1,5-AG was nonnormal and highly right skewed (black histogram). During a median of >21 years of follow-up, there were 1,159 coronary heart disease events, 637 ischemic stroke events, 1,533 heart failure events, and 3,120 deaths. Low values of 1,5-AG (lower than ∼10 μg/mL) were strongly associated with all vascular outcomes and death (Fig. 1). Results were similar when 1,5-AG was modeled using linear splines (Supplementary Fig. 1). We observed a threshold effect, with little evidence of risk associations at (“nondiabetic”) 1,5-AG concentrations of ∼10–15 μg/mL or higher. Indeed, in the categorical analyses, the associations with the clinical outcomes were largely confined to persons with diagnosed diabetes (Table 2 and Supplementary Table 1). Among persons with diagnosed diabetes, those with 1,5-AG <6 μg/mL had a significantly increased risk of coronary heart disease, ischemic stroke, heart failure, or death, even after adjustment for traditional diabetes and cardiovascular risk factors (Table 2 [model 2]). The associations of low 1,5-AG with the coronary heart disease, heart failure, and death were attenuated but remained significant even after further adjustment for HbA1c (Table 2 [model 3]) or fasting glucose (Table 2 [model 4]). The association with ischemic stroke remained significant after additional adjustment for fasting glucose (Table 2 [model 4]) but not HbA1c (Table 2 [model 3]). We did not observe interactions by sex or race for any of the outcomes, but there was some evidence for modest effect modification by age for risk of heart failure and death (Supplementary Table 2). The associations of 1,5-AG with heart failure and death were somewhat stronger in younger persons (<57 years of age) compared with older persons (≥57 years of age).

Figure 1

Adjusted HRs (95% CI) for baseline 1,5-AG with incident coronary heart disease, ischemic stroke, heart failure, and all-cause mortality: the ARIC study, N = 11,106. Adjusted HRs are from Cox proportional hazard regression models. 1,5-AG was modeled using restricted cubic splines (solid line) with knots at the 5th, 35th, 65th, and 95th percentiles and centered at the 10th percentile. 95% CIs are shown with the dashed lines. Models are adjusted for systolic blood pressure (mmHg), blood pressure–lowering medication use (yes, no), education (less than high school, high school or equivalent, more than high school), drinking status (current, former, never), smoking status (current, former, never), physical activity index, and estimated glomerular filtration rate (mL/min per 1.73 m2, modeled using a linear spline with a knot at the median). Frequency histograms are shown for persons without diabetes (light gray bars) and for persons with diabetes (black bars).

Figure 1

Adjusted HRs (95% CI) for baseline 1,5-AG with incident coronary heart disease, ischemic stroke, heart failure, and all-cause mortality: the ARIC study, N = 11,106. Adjusted HRs are from Cox proportional hazard regression models. 1,5-AG was modeled using restricted cubic splines (solid line) with knots at the 5th, 35th, 65th, and 95th percentiles and centered at the 10th percentile. 95% CIs are shown with the dashed lines. Models are adjusted for systolic blood pressure (mmHg), blood pressure–lowering medication use (yes, no), education (less than high school, high school or equivalent, more than high school), drinking status (current, former, never), smoking status (current, former, never), physical activity index, and estimated glomerular filtration rate (mL/min per 1.73 m2, modeled using a linear spline with a knot at the median). Frequency histograms are shown for persons without diabetes (light gray bars) and for persons with diabetes (black bars).

Close modal
Table 2

Adjusted HRs (95% CI) of baseline diabetes-specific categories of 1,5-AG with incident coronary heart disease, ischemic stroke, heart failure, and mortality

OutcomeModel 1 Model 2 Model 3 Model 4
Coronary heart disease (N = 1,159 events)     
 No diagnosis of diabetes     
  1,5-AG ≥6 μg/mL 1.00 (reference) 1.00 (reference) 1.00 (reference) 1.00 (reference) 
  1,5-AG <6 μg/mL 1.19 (0.77, 1.84) 1.12 (0.72, 1.72) 0.80 (0.50, 1.27) 0.93 (0.59, 1.46) 
 Diagnosed diabetes     
  1,5-AG ≥6 μg/mL 2.28 (1.81, 2.87) 1.86 (1.47, 2.36) 1.57 (1.23, 2.01) 1.66 (1.29, 2.12) 
  1,5-AG <6 μg/mL 4.48 (3.66, 5.49)* 3.85 (3.11, 4.78)* 1.86 (1.27, 2.74) 2.47 (1.71, 3.57)* 
  P for trend <0.0001 <0.0001 <0.0001 <0.0001 
Ischemic stroke (N = 637 events)     
 No diagnosis of diabetes     
  1,5-AG ≥6 μg/mL 1.00 (reference) 1.00 (reference) 1.00 (reference) 1.00 (reference) 
  1,5-AG <6 μg/mL 2.79 (1.92, 4.06)* 2.43 (1.66, 3.55)* 1.53 (0.98, 2.39) 2.29 (1.52, 3.45)* 
 Diagnosed diabetes     
  1,5-AG ≥6 μg/mL 1.34 (0.92, 1.95) 1.12 (0.77, 1.64) 0.92 (0.63, 1.37) 1.08 (0.73, 1.60) 
  1,5-AG <6 μg/mL 4.12 (3.20, 5.32)* 3.48 (2.66, 4.55)* 1.46 (0.93, 2.29)* 3.03 (1.97, 4.67)* 
   P for overall trend <0.0001 <0.0001 0.3923 0.0001 
Heart failure (N = 1,553 events)     
 No diagnosis of diabetes     
  1,5-AG ≥6 μg/mL 1.00 (reference) 1.00 (reference) 1.00 (reference) 1.00 (reference) 
  1,5-AG <6 μg/mL 1.15 (0.81, 1.65) 0.97 (0.68, 1.39) 0.71 (0.48, 1.05) 0.81 (0.55, 1.18) 
 Diagnosed diabetes     
  1,5-AG ≥6 μg/mL 2.02 (1.65, 2.47) 1.58 (1.29, 1.94) 1.38 (1.12, 1.71) 1.44 (1.16, 1.78) 
  1,5-AG <6 μg/mL 4.37 (3.69, 5.18)* 3.50 (2.93, 4.17)* 1.91 (1.40, 2.60)* 2.44 (1.82, 3.26)* 
   P for overall trend <0.0001 <0.0001 <0.0001 <0.0001 
All-cause mortality (N = 3,120 events)     
 No diagnosis of diabetes     
  1,5-AG ≥6 μg/mL 1.00 (reference) 1.00 (reference) 1.00 (reference) 1.00 (reference) 
  1,5-AG <6 μg/mL 1.47 (1.18, 1.83)* 1.39 (1.12, 1.74)* 1.18 (0.93, 1.50) 1.32 (1.05, 1.67)* 
 Diagnosed diabetes     
  1,5-AG ≥6 μg/mL 1.68 (1.44, 1.95) 1.48 (1.26, 1.72) 1.36 (1.16, 1.59) 1.43 (1.22, 1.68) 
  1,5-AG <6 μg/mL 2.63 (2.28, 3.03)* 2.44 (2.11, 2.83)* 1.66 (1.30, 2.11) 2.16 (1.71, 2.72)* 
   P for overall trend <0.0001 <0.0001 <0.0001 <0.0001 
OutcomeModel 1 Model 2 Model 3 Model 4
Coronary heart disease (N = 1,159 events)     
 No diagnosis of diabetes     
  1,5-AG ≥6 μg/mL 1.00 (reference) 1.00 (reference) 1.00 (reference) 1.00 (reference) 
  1,5-AG <6 μg/mL 1.19 (0.77, 1.84) 1.12 (0.72, 1.72) 0.80 (0.50, 1.27) 0.93 (0.59, 1.46) 
 Diagnosed diabetes     
  1,5-AG ≥6 μg/mL 2.28 (1.81, 2.87) 1.86 (1.47, 2.36) 1.57 (1.23, 2.01) 1.66 (1.29, 2.12) 
  1,5-AG <6 μg/mL 4.48 (3.66, 5.49)* 3.85 (3.11, 4.78)* 1.86 (1.27, 2.74) 2.47 (1.71, 3.57)* 
  P for trend <0.0001 <0.0001 <0.0001 <0.0001 
Ischemic stroke (N = 637 events)     
 No diagnosis of diabetes     
  1,5-AG ≥6 μg/mL 1.00 (reference) 1.00 (reference) 1.00 (reference) 1.00 (reference) 
  1,5-AG <6 μg/mL 2.79 (1.92, 4.06)* 2.43 (1.66, 3.55)* 1.53 (0.98, 2.39) 2.29 (1.52, 3.45)* 
 Diagnosed diabetes     
  1,5-AG ≥6 μg/mL 1.34 (0.92, 1.95) 1.12 (0.77, 1.64) 0.92 (0.63, 1.37) 1.08 (0.73, 1.60) 
  1,5-AG <6 μg/mL 4.12 (3.20, 5.32)* 3.48 (2.66, 4.55)* 1.46 (0.93, 2.29)* 3.03 (1.97, 4.67)* 
   P for overall trend <0.0001 <0.0001 0.3923 0.0001 
Heart failure (N = 1,553 events)     
 No diagnosis of diabetes     
  1,5-AG ≥6 μg/mL 1.00 (reference) 1.00 (reference) 1.00 (reference) 1.00 (reference) 
  1,5-AG <6 μg/mL 1.15 (0.81, 1.65) 0.97 (0.68, 1.39) 0.71 (0.48, 1.05) 0.81 (0.55, 1.18) 
 Diagnosed diabetes     
  1,5-AG ≥6 μg/mL 2.02 (1.65, 2.47) 1.58 (1.29, 1.94) 1.38 (1.12, 1.71) 1.44 (1.16, 1.78) 
  1,5-AG <6 μg/mL 4.37 (3.69, 5.18)* 3.50 (2.93, 4.17)* 1.91 (1.40, 2.60)* 2.44 (1.82, 3.26)* 
   P for overall trend <0.0001 <0.0001 <0.0001 <0.0001 
All-cause mortality (N = 3,120 events)     
 No diagnosis of diabetes     
  1,5-AG ≥6 μg/mL 1.00 (reference) 1.00 (reference) 1.00 (reference) 1.00 (reference) 
  1,5-AG <6 μg/mL 1.47 (1.18, 1.83)* 1.39 (1.12, 1.74)* 1.18 (0.93, 1.50) 1.32 (1.05, 1.67)* 
 Diagnosed diabetes     
  1,5-AG ≥6 μg/mL 1.68 (1.44, 1.95) 1.48 (1.26, 1.72) 1.36 (1.16, 1.59) 1.43 (1.22, 1.68) 
  1,5-AG <6 μg/mL 2.63 (2.28, 3.03)* 2.44 (2.11, 2.83)* 1.66 (1.30, 2.11) 2.16 (1.71, 2.72)* 
   P for overall trend <0.0001 <0.0001 <0.0001 <0.0001 

Model 1: age (years), race-center, sex (male, female). Model 2: variables in model 1 plus LDL cholesterol (mg/dL), HDL cholesterol (mg/dL), triglycerides (mg/dL), BMI (kg/m2), waist-to-hip ratio, systolic blood pressure (mmHg), blood pressure–lowering medication use (yes, no), education (less than high school, high school or equivalent, more than high school), drinking status (current, former, never), smoking status (current, former, never), physical activity index, and glomerular filtration rate (mL/min per 1.73 m2, modeled using a linear spline with a knot at the median). Model 3: variables in model 2 plus HbA1c (%). Model 4: variables in model 2 plus fasting glucose (mg/dL). P values for overall trend were calculated by modeling the category medians as a continuous variable.

*Significant (P < 0.05) difference between 1,5-AG categories within diabetes group (no diagnosis of diabetes or diagnosed diabetes).

We found that 1,5-AG, a putative biomarker of hyperglycemic excursions over the prior 1–2 weeks, was strongly associated with cardiovascular outcomes and mortality in the setting of diabetes, even after adjustment for baseline fasting glucose or HbA1c. These data help inform a long-standing debate regarding the independent role of postprandial hyperglycemia as a risk factor for cardiovascular outcomes (17,48,49).

Numerous studies have compared fasting glucose and 2-h glucose as risk factors for cardiovascular events and debated their relative importance. Initial epidemiologic studies suggested that 2-h glucose concentrations were more predictive of cardiovascular outcomes compared with fasting glucose, but some early reports assumed a simple linear association of hyperglycemia with vascular outcomes, and few statistically compared the performance of the different biomarkers of hyperglycemia (15,24). However, there is robust evidence that, in many populations, the association of hyperglycemia with cardiovascular outcomes or mortality is J- or U-shaped (i.e., strongly nonlinear) (30,32,5055). A recent meta-analysis that pooled data from >73 prospective studies including almost 300,000 participants without diagnosed diabetes found J-shaped associations of 2-h glucose, fasting glucose, and HbA1c with cardiovascular outcomes, and it directly challenged the assumption that the 2-h glucose concentrations predict cardiovascular disease better than other measures of hyperglycemia (32). An additional difficulty in the interpretation of the epidemiologic literature relates to uncertainty about how well a single oral glucose tolerance test result captures true underlying disturbances in postprandial glucose metabolism. HbA1c subsumes overall chronic exposure to hyperglycemia during the past ∼2–3 months and thus reflects both pre- and postprandial glucose concentrations. Recent large epidemiologic studies and meta-analyses have further demonstrated that a single HbA1c measurement outperforms either fasting or 2-h glucose for prediction of cardiovascular outcomes and mortality (3032). Because 1,5-AG reflects hyperglycemic excursions over a 1- to 2-week period, evidence for its association with long-term outcomes adds depth to this debate.

There are few epidemiologic data linking 1,5-AG to long-term outcomes (14). Our study adds to the evidence regarding the value of 1,5-AG as a biomarker of hyperglycemic excursion in persons with diabetes. There was a striking threshold effect, with little evidence for any associations at concentrations >10 μg/mL. We found that, at very low concentrations, 1,5-AG adds prognostic value for vascular outcomes and death, even after accounting for traditional biomarkers of hyperglycemia (HbA1c or fasting glucose). Indeed, 1,5-AG may be a useful biomarker to monitor hyperglycemic excursions, but additional studies are needed to understand its possible utility as a tool for diabetes management.

The evidence for an independent contribution of postprandial hyperglycemia to cardiovascular risk has given rise to calls for specifically targeting postprandial hyperglycemia in diabetes management. However, the available clinical trial data informing the value of targeting postprandial glucose to prevent diabetic complications are quite limited (56,57). The Study to Prevent Non-Insulin-Dependent Diabetes Mellitus (STOP-NIDDM) trial evaluated acarbose, an antihyperglycemic drug that decreases postprandial hyperglycemia, and demonstrated a significant reduction in cardiovascular events (a secondary end point in this trial) in the acarbose arm compared with placebo (58). However, the total number of cardiovascular events was very small (n = 15 in the treatment arm and n = 32 in the placebo arm). The ongoing Acarbose Cardiovascular Evaluation (ACE) randomized clinical trial should help inform whether treatment with acarbose and specifically targeting postprandial glucose concentrations can reduce the risk of cardiovascular outcomes (59).

Important limitations of our study include the reliance on a single baseline measurement of 1,5-AG and the lack of information on 2-h postprandial glucose; oral glucose tolerance tests were not performed at the second ARIC examination. There were also fewer ischemic stroke events compared with the other outcomes, with corresponding lower power and less precise results for this outcome, particularly in the categorical analyses. Owing to the observational nature of the study, we are also not able to completely rule out the possibility of residual confounding. Nonetheless, this study was one of the largest community-based epidemiologic analyses of 1,5-AG to date. We had more than two decades of follow-up for the development of adjudicated cardiovascular end points, heart failure hospitalizations, and deaths. Follow-up rates in the ARIC cohort are very high (>90%).

In conclusion, we found that 1,5-AG was strongly and independently associated with cardiovascular outcomes and mortality in persons with a history of diabetes. These data add to the growing evidence for the prognostic value of 1,5-AG for important long-term complications of diabetes.

Acknowledgments. The authors thank the staff and participants of the ARIC study for important contributions.

Funding. This research was supported by National Institutes of Health (NIH)/National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK) grant R01-DK-089174 to E.S. E.S. was also supported by NIH/NIDDK grant K24-DK-106414. A.R. was supported by NIH/National Heart, Lung, and Blood Institute (NHLBI) grant T32-HL-007024. P.L. was supported by NIH/NHLBI grant R01-HL-103706. The ARIC study is carried out as a collaborative study supported by NHLBI contracts (HHSN268201100005C, HHSN268201100006C, HHSN268201100007C, HHSN268201100008C, HHSN268201100009C, HHSN268201100010C, HHSN268201100011C, and HHSN268201100012C).

Duality of Interest. Assays for measurement of 1,5-anhydroglucitol were donated by the GlycoMark Corporation. No other potential conflicts of interest relevant to this article were reported.

Author Contributions. E.S. wrote the manuscript, researched data, and provided funding for the study. A.R. conducted the statistical analyses and reviewed and edited the manuscript. P.L., N.M., and J.S.P. reviewed and edited the manuscript. M.S. oversaw the laboratory measurements and reviewed and edited the manuscript. J.C. reviewed and edited the manuscript and participated in the design of the study. E.S. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Prior Presentation. Parts of this study were presented in abstract form at the 2014 American Heart Association EPI/NPAM Scientific Sessions, San Francisco, CA, 18–21 March 2014.

1.
Standards of medical care in diabetes--2015: summary of revisions
.
Diabetes Care
2015
;
38
(
Suppl.
):
S4
[PubMed]
2.
McGill
JB
,
Cole
TG
,
Nowatzke
W
, et al.;
U.S. trial of the GlycoMark assay
.
Circulating 1,5-anhydroglucitol levels in adult patients with diabetes reflect longitudinal changes of glycemia: a U.S. trial of the GlycoMark assay
.
Diabetes Care
2004
;
27
:
1859
1865
[PubMed]
3.
Seok H, Huh JH, Kim HM, et al. 1,5-anhydroglucitol as a useful marker for assessing short-term glycemic excursions in type 1 diabetes. Diabetes Metab J 2015;39:164–170
4.
Dungan
KM
.
1,5-anhydroglucitol (GlycoMark) as a marker of short-term glycemic control and glycemic excursions
.
Expert Rev Mol Diagn
2008
;
8
:
9
19
[PubMed]
5.
Dungan
KM
,
Buse
JB
,
Largay
J
, et al
.
1,5-anhydroglucitol and postprandial hyperglycemia as measured by continuous glucose monitoring system in moderately controlled patients with diabetes
.
Diabetes Care
2006
;
29
:
1214
1219
[PubMed]
6.
Yamanouchi
T
,
Akanuma
Y
,
Toyota
T
, et al
.
Comparison of 1,5-anhydroglucitol, HbA1c, and fructosamine for detection of diabetes mellitus
.
Diabetes
1991
;
40
:
52
57
[PubMed]
7.
Buse
JB
,
Freeman
JL
,
Edelman
SV
,
Jovanovic
L
,
McGill
JB
.
Serum 1,5-anhydroglucitol (GlycoMark ): a short-term glycemic marker
.
Diabetes Technol Ther
2003
;
5
:
355
363
[PubMed]
8.
Selvin
E
,
Steffes
MW
,
Ballantyne
CM
,
Hoogeveen
RC
,
Coresh
J
,
Brancati
FL
.
Racial differences in glycemic markers: a cross-sectional analysis of community-based data
.
Ann Intern Med
2011
;
154
:
303
309
[PubMed]
9.
Selvin
E
,
Francis
LM
,
Ballantyne
CM
, et al
.
Nontraditional markers of glycemia: associations with microvascular conditions
.
Diabetes Care
2011
;
34
:
960
967
[PubMed]
10.
Juraschek
SP
,
Steffes
MW
,
Miller
ER
 3rd
,
Selvin
E
.
Alternative markers of hyperglycemia and risk of diabetes
.
Diabetes Care
2012
;
35
:
2265
2270
[PubMed]
11.
Juraschek
SP
,
Steffes
MW
,
Selvin
E
.
Associations of alternative markers of glycemia with hemoglobin A(1c) and fasting glucose
.
Clin Chem
2012
;
58
:
1648
1655
[PubMed]
12.
Alssema M, Boers HM, Ceriello A, et al. Diet and glycaemia: the markers and their meaning. A report of the Unilever Nutrition Workshop. Br J Nutr. 11 December 2014 [Epub ahead of print]
[PubMed]
13.
Konya
J
,
Ng
JM
,
Cox
H
, et al
.
Use of complementary markers in assessing glycaemic control in people with diabetic kidney disease undergoing iron or erythropoietin treatment
.
Diabet Med
2013
;
30
:
1250
1254
[PubMed]
14.
Parrinello
CM
,
Selvin
E
.
Beyond HbA1c and glucose: the role of nontraditional glycemic markers in diabetes diagnosis, prognosis, and management
.
Curr Diab Rep
2014
;
14
:
548
[PubMed]
15.
Meigs
JB
,
Nathan
DM
,
D’Agostino
RB
 Sr
,
Wilson
PW
;
Framingham Offspring Study
.
Fasting and postchallenge glycemia and cardiovascular disease risk: the Framingham Offspring Study
.
Diabetes Care
2002
;
25
:
1845
1850
[PubMed]
16.
Lin HJ, Lee BC, Ho YL, et al. Postprandial glucose improves the risk prediction of cardiovascular death beyond the metabolic syndrome in the non-diabetic population. Diabetes Care 2009;32:1721–1726
17.
Cavalot
F
,
Petrelli
A
,
Traversa
M
, et al
.
Postprandial blood glucose is a stronger predictor of cardiovascular events than fasting blood glucose in type 2 diabetes mellitus, particularly in women: lessons from the San Luigi Gonzaga Diabetes Study
.
J Clin Endocrinol Metab
2006
;
91
:
813
819
[PubMed]
18.
Haffner
SM
.
The importance of hyperglycemia in the nonfasting state to the development of cardiovascular disease
.
Endocr Rev
1998
;
19
:
583
592
[PubMed]
19.
Tominaga
M
,
Eguchi
H
,
Manaka
H
,
Igarashi
K
,
Kato
T
,
Sekikawa
A
.
Impaired glucose tolerance is a risk factor for cardiovascular disease, but not impaired fasting glucose. The Funagata Diabetes Study
.
Diabetes Care
1999
;
22
:
920
924
[PubMed]
20.
Bonora
E
,
Muggeo
M
.
Postprandial blood glucose as a risk factor for cardiovascular disease in Type II diabetes: the epidemiological evidence
.
Diabetologia
2001
;
44
:
2107
2114
[PubMed]
21.
Cavalot
F
,
Pagliarino
A
,
Valle
M
, et al
.
Postprandial blood glucose predicts cardiovascular events and all-cause mortality in type 2 diabetes in a 14-year follow-up: lessons from the San Luigi Gonzaga Diabetes Study
.
Diabetes Care
2011
;
34
:
2237
2243
[PubMed]
22.
Yudkin
JS
.
Post-load hyperglycaemia-an inappropriate therapeutic target
.
Lancet
2002
;
359
:
166
167
[PubMed]
23.
Donahue
RP
,
Abbott
RD
,
Reed
DM
,
Yano
K
.
Postchallenge glucose concentration and coronary heart disease in men of Japanese ancestry. Honolulu Heart Program
.
Diabetes
1987
;
36
:
689
692
[PubMed]
24.
Glucose tolerance and mortality: comparison of WHO and American Diabetes Association diagnostic criteria. The DECODE study group. European Diabetes Epidemiology Group. Diabetes Epidemiology: Collaborative analysis Of Diagnostic criteria in Europe
.
Lancet
1999
;
354
:
617
621
[PubMed]
25.
Davidson
MB
.
Counterpoint: postprandial glucose levels are not a clinically important treatment target
.
Diabetes Care
2010
;
33
:
1908
1910
[PubMed]
26.
Ceriello
A
,
Esposito
K
,
Piconi
L
, et al
.
Oscillating glucose is more deleterious to endothelial function and oxidative stress than mean glucose in normal and type 2 diabetic patients
.
Diabetes
2008
;
57
:
1349
1354
[PubMed]
27.
Wascher
TC
,
Schmoelzer
I
,
Wiegratz
A
, et al
.
Reduction of postchallenge hyperglycaemia prevents acute endothelial dysfunction in subjects with impaired glucose tolerance
.
Eur J Clin Invest
2005
;
35
:
551
557
[PubMed]
28.
Coutinho
M
,
Gerstein
HC
,
Wang
Y
,
Yusuf
S
.
The relationship between glucose and incident cardiovascular events. A metaregression analysis of published data from 20 studies of 95,783 individuals followed for 12.4 years
.
Diabetes Care
1999
;
22
:
233
240
[PubMed]
29.
Pankow
JS
,
Kwan
DK
,
Duncan
BB
, et al
.
Cardiometabolic risk in impaired fasting glucose and impaired glucose tolerance: the Atherosclerosis Risk in Communities Study
.
Diabetes Care
2007
;
30
:
325
331
[PubMed]
30.
Selvin
E
,
Steffes
MW
,
Zhu
H
, et al
.
Glycated hemoglobin, diabetes, and cardiovascular risk in nondiabetic adults
.
N Engl J Med
2010
;
362
:
800
811
[PubMed]
31.
Sarwar
N
,
Aspelund
T
,
Eiriksdottir
G
, et al
.
Markers of dysglycaemia and risk of coronary heart disease in people without diabetes: Reykjavik prospective study and systematic review
.
PLoS Med
2010
;
7
:
e1000278
[PubMed]
32.
Di Angelantonio
E
,
Gao
P
,
Khan
H
, et al.;
Emerging Risk Factors Collaboration
.
Glycated hemoglobin measurement and prediction of cardiovascular disease
.
JAMA
2014
;
311
:
1225
1233
[PubMed]
33.
Esposito
K
,
Giugliano
D
,
Nappo
F
,
Marfella
R
;
Campanian Postprandial Hyperglycemia Study Group
.
Regression of carotid atherosclerosis by control of postprandial hyperglycemia in type 2 diabetes mellitus
.
Circulation
2004
;
110
:
214
219
[PubMed]
34.
Selvin
E
,
Rawlings
AM
,
Grams
M
,
Klein
R
,
Steffes
M
,
Coresh
J
.
Association of 1,5-anhydroglucitol with diabetes and microvascular conditions
.
Clin Chem
2014
;
60
:
1409
1418
[PubMed]
35.
The Atherosclerosis Risk in Communities (ARIC) Study: design and objectives. The ARIC investigators
.
Am J Epidemiol
1989
;
129
:
687
702
[PubMed]
36.
Selvin
E
,
Rynders
GP
,
Steffes
MW
.
Comparison of two assays for serum 1,5-anhydroglucitol
.
Clin Chim Acta
2011
;
412
:
793
795
[PubMed]
37.
Selvin
E
,
Coresh
J
,
Zhu
H
,
Folsom
A
,
Steffes
MW
.
Measurement of HbA1c from stored whole blood samples in the Atherosclerosis Risk in Communities study
.
J Diabetes
2010
;
2
:
118
124
[PubMed]
38.
Siedel
J
,
Hägele
EO
,
Ziegenhorn
J
,
Wahlefeld
AW
.
Reagent for the enzymatic determination of serum total cholesterol with improved lipolytic efficiency
.
Clin Chem
1983
;
29
:
1075
1080
[PubMed]
39.
Friedewald
WT
,
Levy
RI
,
Fredrickson
DS
.
Estimation of the concentration of low-density lipoprotein cholesterol in plasma, without use of the preparative ultracentrifuge
.
Clin Chem
1972
;
18
:
499
502
[PubMed]
40.
Nägele
U
,
Hägele
EO
,
Sauer
G
, et al
.
Reagent for the enzymatic determination of serum total triglycerides with improved lipolytic efficiency
.
J Clin Chem Clin Biochem
1984
;
22
:
165
174
[PubMed]
41.
Operations Manual No. 10: Clinical Chemistry Determinations, Version 1.0. Chapel Hill, NC, University of North Carolina, 1987
42.
Operations Manual No. 2: Cohort Component Procedures, Version 1.0. Chapel Hill, NC, University of North Carolina, 1987
43.
Levey
AS
,
Stevens
LA
,
Schmid
CH
, et al.;
CKD-EPI (Chronic Kidney Disease Epidemiology Collaboration)
.
A new equation to estimate glomerular filtration rate
.
Ann Intern Med
2009
;
150
:
604
612
[PubMed]
44.
Baecke
JA
,
Burema
J
,
Frijters
JE
.
A short questionnaire for the measurement of habitual physical activity in epidemiological studies
.
Am J Clin Nutr
1982
;
36
:
936
942
[PubMed]
45.
Rosamond
WD
,
Folsom
AR
,
Chambless
LE
, et al
.
Stroke incidence and survival among middle-aged adults: 9-year follow-up of the Atherosclerosis Risk in Communities (ARIC) cohort
.
Stroke
1999
;
30
:
736
743
[PubMed]
46.
White
AD
,
Folsom
AR
,
Chambless
LE
, et al
.
Community surveillance of coronary heart disease in the Atherosclerosis Risk in Communities (ARIC) study: methods and initial two years’ experience
.
J Clin Epidemiol
1996
;
49
:
223
233
[PubMed]
47.
Harrell
FE
.
Regression Modeling Strategies With Applications to Linear Models, Logistic Regression, and Survival Analysis
.
New York
,
Springer
,
2001
48.
Nathan
DM
,
Davidson
MB
,
DeFronzo
RA
, et al.;
American Diabetes Association
.
Impaired fasting glucose and impaired glucose tolerance: implications for care
.
Diabetes Care
2007
;
30
:
753
759
[PubMed]
49.
American Diabetes Association
.
Postprandial blood glucose
.
Diabetes Care
2001
;
24
:
775
778
[PubMed]
50.
Paprott
R
,
Schaffrath Rosario
A
,
Busch
MA
, et al
.
Association between hemoglobin A1c and all-cause mortality: results of the mortality follow-up of the German National Health Interview and Examination Survey 1998
.
Diabetes Care
2015
;
38
:
249
256
[PubMed]
51.
Saydah
SH
,
Miret
M
,
Sung
J
,
Varas
C
,
Gause
D
,
Brancati
FL
.
Postchallenge hyperglycemia and mortality in a national sample of U.S. adults
.
Diabetes Care
2001
;
24
:
1397
1402
[PubMed]
52.
Brewer
N
,
Wright
CS
,
Travier
N
, et al
.
A New Zealand linkage study examining the associations between A1C concentration and mortality
.
Diabetes Care
2008
;
31
:
1144
1149
[PubMed]
53.
Barr
EL
,
Boyko
EJ
,
Zimmet
PZ
,
Wolfe
R
,
Tonkin
AM
,
Shaw
JE
.
Continuous relationships between non-diabetic hyperglycaemia and both cardiovascular disease and all-cause mortality: the Australian Diabetes, Obesity, and Lifestyle (AusDiab) study
.
Diabetologia
2009
;
52
:
415
424
[PubMed]
54.
Balkau
B
,
Bertrais
S
,
Ducimetiere
P
,
Eschwege
E
.
Is there a glycemic threshold for mortality risk
?
Diabetes Care
1999
;
22
:
696
699
[PubMed]
55.
DECODE Study Group, European Diabetes Epidemiology Group
.
Is the current definition for diabetes relevant to mortality risk from all causes and cardiovascular and noncardiovascular diseases
?
Diabetes Care
2003
;
26
:
688
696
[PubMed]
56.
Raz
I
,
Ceriello
A
,
Wilson
PW
, et al
.
Post hoc subgroup analysis of the HEART2D trial demonstrates lower cardiovascular risk in older patients targeting postprandial versus fasting/premeal glycemia
.
Diabetes Care
2011
;
34
:
1511
1513
[PubMed]
57.
Raz
I
,
Wilson
PW
,
Strojek
K
, et al
.
Effects of prandial versus fasting glycemia on cardiovascular outcomes in type 2 diabetes: the HEART2D trial
.
Diabetes Care
2009
;
32
:
381
386
[PubMed]
58.
Chiasson
JL
,
Josse
RG
,
Gomis
R
,
Hanefeld
M
,
Karasik
A
,
Laakso
M
;
STOP-NIDDM Trial Research Group
.
Acarbose treatment and the risk of cardiovascular disease and hypertension in patients with impaired glucose tolerance: the STOP-NIDDM trial
.
JAMA
2003
;
290
:
486
494
[PubMed]
59.
Holman RR, Bethel MA, Chan JC, et al. Rationale for and design of the acarbose cardiovascular evaluation (ace) trial. Am Heart J 2014;168:23–29.e22

Supplementary data