Abstract
BACKGROUND: 1,5-Anhydroglucitol (1,5-AG) is inversely related to hyperglycemia and may be a useful indicator of short-term (1–2 weeks) hyperglycemia and glycemic excursions, but its prognostic value is unclear. We sought to evaluate the associations of 1,5-AG with risk of diabetes and microvascular disease.
METHODS: We measured 1,5-AG in blood samples from over 10 000 participants in the ARIC (Atherosclerosis Risk in Communities) Study (baseline, 1990–1992) and characterized the independent associations with prevalent retinopathy and with incident chronic kidney disease and incident diabetes during approximately 20 years of follow-up.
RESULTS: We found that 1,5-AG was associated with prevalent retinopathy, driven primarily by the strong association in persons with diagnosed diabetes: adjusted odds ratio (OR) 11.26 (95% CI, 6.17–20.53) for <6 μg/mL compared to 1,5-AG ≥10 μg/mL. This result remained significant after further adjustment for hemoglobin A1c (Hb A1c) (OR, 4.85; 95% CI, 2.42–9.74). In persons with diagnosed diabetes, low 1,5-AG (<6 μg/mL vs ≥10 μg/mL) was also associated with a >2-fold increased risk of incident chronic kidney disease [adjusted hazard ratio (HR), 2.83; 95% CI, 2.15–3.74] and remained significant after adjustment for Hb A1c (HR, 1.43; 95% CI, 1.02–2.00). Nondiabetic participants with high 1,5-AG (≥10 μg/mL) had the lowest prevalence of retinopathy and lowest risk of kidney disease. In persons without diagnosed diabetes at baseline, 1,5-AG <10 μg/mL was also associated with incident diabetes (adjusted HR, 2.29; 95% CI, 2.03–2.58).
CONCLUSIONS: 1,5-AG was associated with long-term risk of important microvascular outcomes, particularly in persons with diagnosed diabetes and even after adjustment for Hb A1c. Our results suggest 1,5-AG may capture risk information associated with hyperglycemic excursions.
Hemoglobin A1c (Hb A1c)6 is the standard clinical measure used to monitor glycemic control, and it is now recommended for use in the diagnosis of diabetes (1). Although Hb A1c has high reliability compared to the more traditional oral glucose tolerance test (2), there are certain settings in which Hb A1c testing can be problematic (e.g., anemias, hemoglobinopathies, dialysis, pregnancy, liver disease). In addition, there is evidence that glycemic excursions (an aspect of diabetes control incompletely captured by Hb A1c) may contribute to vascular damage independently of mean glucose concentrations (3–5). As such, there is growing interest in alternative markers of hyperglycemia.
1,5-Anhydroglucitol (1,5-AG) or 1-deoxyglucose is a monosaccharide, originating mainly from foods and closely resembling glucose in structure (6, 7). In the normoglycemic setting, 1,5-AG is typically present at high but constant concentrations in the blood. 1,5-AG is freely filtered by the glomeruli and reabsorbed in the renal tubule with the small amount excreted corresponding to dietary intake. However, in the setting of hyperglycemia, high amounts of glucose block tubular reabsorption of 1,5-AG, causing serum concentrations to fall (i.e., 1,5-AG has an inverse association with serum glucose). Thus, 1,5-AG is thought to be an indicator of postprandial glucose excursions and short-term (1–2 weeks) hyperglycemia (8, 9).
Although the assay for 1,5-AG is commercially available, there are few data linking concentrations to long-term outcomes. It is unclear if 1,5-AG is associated with microvascular complications of diabetes and if 1,5-AG adds prognostic value to Hb A1c. We conducted this study to characterize the association of 1,5-AG with prevalent retinopathy, incident chronic kidney disease (CKD), and incident diabetes.
Methods
STUDY POPULATION
The Atherosclerosis Risk in Communities (ARIC) Study is a community-based prospective cohort of over 15 000 participants sampled from 4 US communities. The first clinic examinations (visit 1) took place from 1987 to 1989, with 3 follow-up visits approximately every 3 years (10). A fifth visit was completed in 2011–2013. The second clinic examination (visit 2) took place from 1990 to 1992 and is the baseline for the present study. There were 14 348 participants who attended visit 2. Institutional review boards at each clinical site reviewed the study and informed consent was obtained from all participants.
In the present study, we excluded all persons whose race/ethnicity was recorded as other than white or black (n = 42), who were fasting <8 h (n = 446), or who were missing variables of interest (n = 1552). Retinal photographs were not taken at baseline. Thus, for our analyses of retinopathy, we included only those participants who met our inclusion criteria and who also attended visit 3 (1993–1995), at which time retinal photographs were taken (final analytic sample, n = 9447). For analyses of incident CKD, we excluded persons with reduced kidney function, defined as an estimated glomerular filtration rate (eGFR) [calculated from serum creatinine using the 2009 CKD–Epidemiology Collaboration (CKD-EPI) equation] of <60 mL · min−1 · (1.73 m2)−1 at baseline (final analytic sample, n = 12 083). For analyses of incident diabetes, we excluded persons with diagnosed diabetes at baseline (final analytic sample, n = 10 948).
MEASUREMENT OF 1,5-AG
1,5-AG (GlycoMark) was measured in 2012–2013 in stored serum samples from visit 2 (1990–1992) using a Roche Modular P800 system. The interassay CV was 5%. The reliability coefficient for n = 610 masked duplicate sample pairs was 0.99. Previous studies have shown this 1,5-AG assay to be reliable in long-term stored samples (11, 12).
ASSESSMENT OF RETINOPATHY
Retinal photographs were taken at visit 3 (1993–1995) following a standardized protocol that has been previously described (13, 14). Briefly, after 5 minutes of dark adaptation, a nonmydriatic 45-degree retinal photograph centered on the optic disc and macula was taken of 1 randomly selected eye. Trained graders masked to participant information evaluated each of the photographs. We defined any retinopathy as a severity score ≥20 according to a modification of the Airlie House classification system, as used in the modified Early Treatment Diabetic Retinopathy Study (ETDRS) (13). Level 20 is commonly considered the earliest stage of diabetic retinopathy (15).
ASSESSMENT OF INCIDENT CHRONIC KIDNEY DISEASE
We used the 2009 CKD-EPI creatinine equation to determine the eGFR. Among persons with normal kidney function at baseline (visit 2), we defined incident CKD as either eGFR <60 mL · min−1 · (1.73 m2)−1 estimated from serum creatinine measured at visit 4 (1996–1998), accompanied by at least a 25% decrease in eGFR from baseline (visit 2), or a kidney disease hospitalization or death identified during continuous active surveillance (16).
ASSESSMENT OF INCIDENT DIABETES
We identified incident (new) cases of diabetes on the basis of a self-reported diabetes diagnosis or use of diabetes medications during the ARIC visits and subsequent annual telephone calls with follow-up through April 2011. In sensitivity analyses, we evaluated 2 additional definitions of incident diabetes incorporating information on undiagnosed cases identified using glucose measurements at 2 subsequent ARIC visits (17).
OTHER VARIABLES
Serum glucose was measured using the hexokinase method. Hb A1c was measured in stored whole blood samples using HPLC with instruments standardized to the Diabetes Control and Complications Trial assay (Tosoh A1c 2.2 Plus Glycohemoglobin Analyzer and Tosoh G7) (18). Plasma lipid concentrations (19–22), body mass index (23), and blood pressure (24) were also measured. 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. Participants reported their education level, alcohol use, and smoking status. The level of physical activity was assessed with Baecke's questionnaire at visit 1 (25).
STATISTICAL ANALYSES
Baseline characteristics of the study population were calculated overall and by categories of 1,5-AG at baseline. We categorized 1,5-AG according to cutpoints recommended by the manufacturer. In persons with a diagnosis of diabetes, we divided the population into 3 groups on the basis of concentrations of 1,5-AG at baseline (<6, 6 to <10, and ≥10 μg/mL). In persons without a diagnosis of diabetes, we divided the population into 2 groups on the basis of concentrations of <10 and ≥10 μg/mL.
Adjusted odds ratios (ORs) and their corresponding 95% CIs for retinopathy were estimated using multivariable logistic regression models. Model accuracy was assessed using the area under the ROC curve (AUC). For analyses of incident CKD and diabetes, adjusted hazard ratios (HRs) and their corresponding 95% CIs were estimated using Cox proportional hazards models. We verified that the proportional hazards assumption was met using log–log plots and by testing for risk factor-by-time interactions. To characterize the shape and assess the continuous associations of 1,5-AG with each of the clinical outcomes, we fit restricted cubic splines (26). Model discrimination was assessed using Harrell's c-statistic (27). To evaluate the overall improvement in risk classification for the addition of 1,5-AG to adjusted models (including models with Hb A1c or fasting glucose), we calculated the continuous net-reclassification improvement (NRI) statistic and the integrated-discrimination-improvement (IDI) statistic (28).
We constructed 4 models for each of the outcomes. Model 1 was adjusted for age, sex, race-center (5 categories: 3 categories being white participants in Minnesota, Maryland, and North Carolina; and 2 categories being black participants in North Carolina and Mississippi). Model 2 was adjusted for all variables in model 1 plus LDL cholesterol, HDL cholesterol, triglycerides, body mass index, waist-to-hip ratio, systolic blood pressure, blood pressure–lowering medication use, family history of diabetes, education level, alcohol use, smoking status, and physical activity level. Model 3 was adjusted for all variables in model 2 plus Hb A1c. Model 4 was adjusted for all variables in model 2 plus fasting glucose. In analyses comparing categories of 1,5-AG at baseline, persons with diagnosed diabetes and 1,5-AG concentrations ≥10 μg/mL served as the reference group. We tested for interactions by sex and race.
To evaluate whether 1,5-AG added prognostic information within clinically relevant categories of Hb A1c among persons with diagnosed diabetes, we calculated the crude prevalence of retinopathy and 20-year cumulative probability (Kaplan–Meier method) of CKD by categories of 1,5-AG (<10 and ≥10 μg/mL) within categories of glycemic control at baseline [Hb A1c, <7% and ≥7% (53 mmol/mol)]. We compared the prevalence estimates using the Z-test and the 20-year probabilities using the log-rank test.
All analyses were conducted using Stata/SE version 13.0 (StataCorp).
Results
Baseline characteristics of the study population according to categories of 1,5-AG in persons with and without diagnosed diabetes are shown in Table 1. In general, baseline risk factor associations for 1,5-AG were inverse to but highly consistent with known risk factors of diabetes and hyperglycemia. Older age, black race, higher body mass index, hypertension, lower education, higher proportion family history of diabetes, lower HDL cholesterol, and higher triglycerides were all associated with lower baseline concentrations of 1,5-AG. 1,5-AG was also strongly inversely associated with both Hb A1c and fasting glucose, although the associations were nonlinear, with a flat association between 1,5-AG, Hb A1c and fasting glucose at 1,5-AG concentrations >10 μg/mL (see Fig. 1 in the Data Supplement that accompanies the online version of this article at http://www.clinchem.org/content/vol60/issue11). In persons with diagnosed diabetes, the Spearman correlations with 1,5-AG were −0.84 for Hb A1c and −0.77 for fasting glucose. Among persons without diagnosed diabetes the correlations of Hb A1c and fasting glucose with 1,5-AG were both low (Spearman correlations of −0.08 and −0.02 for Hb A1c and fasting glucose, respectively) (see online Supplemental Fig. 1). In persons with diagnosed diabetes, the 1,5-AG cut-point values of 10 μg/mL and 6 μg/mL represented the 64th and 50th percentiles, respectively, corresponding to fasting glucose values of approximately 165 mg/dL and 194 mg/dL (9.16 and 10.77 mmol/L) and Hb A1c values of 7.3% and 8.2% (see online Supplemental Table 1). Spearman correlations between 1,5-AG and the other continuous variables were generally low regardless of diabetes status (see online Supplemental Fig. 2).
In the 9447 persons with retinal photographs available, there were 332 cases of retinopathy. In persons with diagnosed diabetes, low concentrations of 1,5-AG were strongly associated with retinopathy (Table 2). The lowest prevalence of retinopathy was in persons without diagnosed diabetes and 1,5-AG of ≥10 μg/mL, but this association was attenuated after adjustment for major risk factors and especially after further adjustment for Hb A1c or fasting glucose. In persons with diagnosed diabetes, those with 1,5-AG <6 μg/mL were 11 times more likely to have retinopathy at visit 3 compared to persons with diagnosed diabetes and 1,5-AG of ≥10 μg/mL even after adjustment (model 2, OR, 11.26; 95% CI, 6.17–20.53) (Table 2). These results were attenuated but remained significant after adjustment for Hb A1c or fasting glucose (Table 2).
In the 12 083 persons with normal kidney function at baseline, there were 1534 incident cases of CKD during a median of 19 years of follow-up. 1,5-AG was strongly associated with incident CKD in a largely graded fashion. Persons without diagnosed diabetes and with high 1,5-AG concentrations (≥10 μg/mL) had the lowest risk of incident CKD, whereas, persons with diagnosed diabetes and 1,5-AG concentrations <6 μg/mL had almost a 3-fold increased risk of CKD even after multivariable adjustment (Table 2, Model 2: HR, 2.83; 95% CI, 2.15–3.74). After adjustment for Hb A1c or fasting glucose, the association between lower 1,5-AG concentrations and incident CKD in persons with diabetes was attenuated but remained significant (Table 2, Models 3 and 4).
In the 10 948 persons without a diagnosis of diabetes at baseline, there were 2882 incident cases of diabetes during a median of 18 years of follow-up. Over this time period, persons with 1,5-AG concentrations of <10 μg/mL at baseline were more than twice as likely to develop diabetes compared to persons with 1,5-AG ≥10 μg/mL. The association persisted after comprehensive adjustment for traditional diabetes risk factors (Table 2, Model 2: HR, 2.29 (95% CI, 2.03–2.58) and was attenuated after further adjustment for Hb A1c (Model 3: 1.42; 95% CI, 1.24–1.63) or fasting glucose (Model 4: 1.22; 95% CI, 1.05–1.43). Sensitivity analyses demonstrated that these results were robust to different definitions of incident diabetes (see online Supplemental Table 2).
We did not observe statistically significant interactions by race for 1,5-AG with any of the outcomes. There were also no statistically significant interactions by sex for 1,5-AG with retinopathy or CKD. We did, however, observe a somewhat stronger association of 1,5-AG ≥10 μg/mL (vs <10 μg/mL) with incident diabetes in men (Model 2: HR, 3.05; 95% CI, 2.56–3.63) compared to women (Model 2: HR, 1.83; 95% CI, 1.56–2.15).
1,5-AG significantly improved prediction of retinopathy, CKD, and diabetes when added to a model with basic risk factors (Table 3). When 1,5-AG was added to models already containing Hb A1c, the improvement in the c-statistic was statistically significant for retinopathy but not incident CKD or incident diabetes. The continuous NRI and IDI statistics for retinopathy, incident CKD, and incident diabetes showed largely similar results (see online Supplemental Table 3).
In the overall population (persons with and without diabetes combined), the relative association of 1,5-AG was much stronger for retinopathy compared to incident CKD, although the shapes of the associations of 1,5-AG were similar (Fig. 1). The associations of 1,5-AG with prevalent retinopathy, incident CKD, and incident diabetes appeared strong and linear at values below approximately 15 μg/mL, but there was little evidence for associations at higher values of 1,5-AG. The substantially different distributions of 1,5-AG in persons with and without diabetes are evident from the frequency histograms included in Fig. 1. Consistent with the results presented in Table 2, analyses of prevalent retinopathy and incident CKD stratified by diabetes status at baseline reveal that much of the association is being driven by persons with diabetes and/or low concentrations of 1,5-AG at baseline (see online Supplemental Fig. 3).
Frequency histograms for 1,5-AG are shown separately for persons with diagnosed diabetes (dark grey bars) and without diagnosed diabetes (light grey bars). Adjusted ORs (for prevalent retinopathy) are from logistic regression models and adjusted HRs (for incident CKD and incident diabetes) are from Cox proportional hazards models. Baseline 1,5-AG was modeled using restricted cubic splines (solid lines) with knots at the 5th, 35th, 65th, and 95th percentiles. Models were centered at the 10th percentile of 1,5-AG. Display of the data was truncated at the 1st and 99th percentiles. The shaded areas are the 95% CIs for restricted cubic spline models. Models were adjusted for age (years), race-center, sex (male, female), LDL cholesterol (mg/dL), HDL cholesterol (mg/dL), triglycerides (mg/dL), body mass index (kg/m2), waist-to-hip ratio, mean systolic blood pressure (mmHg), blood pressure–lowering medication use (yes, no), family history of diabetes (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), and physical activity index (score).
Among persons with a diagnosis of diabetes, the prevalence of retinopathy and absolute risk of CKD associated with categories of 1,5-AG (<10 and ≥10 μg/mL) by categories of Hb A1c [<7 and ≥7%) (<53 and ≥53 mmol/mol)] are shown in Fig. 2. Among persons with Hb A1c ≥7% (≥53 mmol/mol), persons with 1,5-AG <10 μg/mL had substantially more retinopathy than those with 1,5-AG concentrations ≥10 μg/mL (P < 0.001) (Fig. 2A). Among persons with Hb A1c <7% (<53 mmol/mol), 1,5-AG categories did not provide significant additional risk stratification information (P value = 0.10). A similar pattern was observed for the cumulative incidence of CKD (Fig. 2B).
Conclusions
This study demonstrated strong associations between 1,5-AG and diabetic microvascular outcomes in a community-based setting and suggests that 1,5-AG has prognostic value in persons with diabetes and particularly those with Hb A1c ≥7% (≥53 mmol/mol). Low concentrations of 1,5-AG were strongly associated with prevalent retinopathy and incident CKD, particularly among those persons with diagnosed diabetes. In addition, 1,5-AG added significant information beyond Hb A1c for discrimination of prevalent retinopathy. These results suggest that 1,5-AG may complement Hb A1c for monitoring recent postprandial glycemic excursions in persons with diabetes, but additional studies are needed to establish its clinical utility.
In persons without diabetes, the association of 1,5-AG with the development of diabetes and complications of diabetes was moderate to weak. Indeed, among persons without diabetes, the correlations of 1,5-AG with fasting glucose and Hb A1c were very low. These findings are in contrast to those for other markers of hyperglycemia, such as glycated albumin and fructosamine, which are more strongly associated with Hb A1c and fasting glucose (29), and predict long-term outcomes even in persons without a diagnosis of diabetes (30). There was no evidence of associations of 1,5-AG with long-term outcomes at values of approximately 15 μg/mL or higher. This suggests that while low concentrations of 1,5-AG are relatively insensitive for the identification of early hyperglycemic states and future diabetes, they may be useful in the setting of overt diabetes. Our findings are consistent with the physiology of 1,5-AG, in which plasma concentrations of 1,5-AG are thought to show decreases only at the highest concentrations of blood glucose and reflect glucose excursions (31–33).
1,5-AG has been shown in other studies to correlate with complications of diabetes. In a community-based cross-sectional study of 517 persons with type 2 diabetes in Korea, low 1,5-AG concentrations were associated with both retinopathy and albuminuria (34). We have previously shown cross-sectional associations of 1,5-AG with retinopathy and albuminuria in 1600 older adults in ARIC (mean age, 70 years) (35). In one previous prospective study of approximately 2000 persons in Japan, 1,5-AG at baseline was significantly associated with incident cardiovascular events during 11 years of follow-up (36). Our results extend the findings reported in this literature and suggest that 1,5-AG may be a useful biomarker of prognosis for microvascular outcomes in the setting of diabetes.
Nonetheless, certain limitations of our study should be considered in the interpretation of these data. These include the reliance on a single measurement of 1,5-AG, the limited number of fasting glucose and serum creatinine measurements during follow-up, and the fact that the retinal photographs were obtained only at visit 3 (in 1993–1994), whereas measurements of 1,5-AG were obtained at baseline (in 1990–1992). We were not able to validate the incident self-reported cases of diabetes identified after visit 4. Nonetheless, in sensitivity analyses of incident diabetes defined on the basis of a combination of glucose measurement, medication use, and self-reported information available for the first 6 years of follow-up, our results were similar. Strengths of this study include the large, community-based sample, the rigorous measurement of diabetes risk factors, and the long-term prospective follow-up with high retention (>90% contact rate during follow-up in ARIC).
In summary, this study demonstrated robust associations between low concentrations of 1,5-AG and microvascular complications in the setting of diabetes and supports a possible role for 1,5-AG as a useful biomarker of hyperglycemia. Additional studies are needed to fully evaluate the clinical utility of 1,5-AG in the setting of diabetes management.
Acknowledgments
The authors thank the staff and participants of the ARIC study for their important contributions. Dr. E. Selvin 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.
Footnotes
↵6 Nonstandard abbreviations:
- Hb A1c,
- hemoglobin A1c;
- 1,5-AG,
- 1,5-anhydroglucitol;
- CKD,
- chronic kidney disease;
- ARIC,
- Atherosclerosis Risk in Communities;
- eGFR,
- estimated glomerular filtration rate;
- CKD-EPI,
- CKD–Epidemiology Collaboration;
- ETDRS,
- Early Treatment Diabetic Retinopathy Study;
- OR,
- odds ratio;
- AUC,
- area under the ROC curve;
- HR,
- hazard ratio;
- NRI,
- net-reclassification improvement;
- IDI,
- integrated-discrimination-improvement.
(see editorial on page 1359)
Author Contributions: All authors confirmed they have contributed to the intellectual content of this paper and have met the following 3 requirements: (a) significant contributions to the conception and design, acquisition of data, or analysis and interpretation of data; (b) drafting or revising the article for intellectual content; and (c) final approval of the published article.
Authors' Disclosures or Potential Conflicts of Interest: Upon manuscript submission, all authors completed the author disclosure form. Disclosures and/or potential conflicts of interest:
Employment or Leadership: None declared.
Consultant or Advisory Role: None declared.
Stock Ownership: None declared.
Honoraria: None declared.
Research Funding: The Atherosclerosis Risk in Communities Study is carried out as a collaborative study supported by National Heart, Lung, and Blood Institute contracts (HHSN268201100005C, HHSN268201100006C, HHSN268201100007C, HHSN268201100008C, HHSN268201100009C, HHSN268201100010C, HHSN268201100011C, and HHSN268201100012C) and reagents for the 1,5-AG assays were donated by the GlycoMarkTM Corporation; E. Selvin, NIH/NIDDK grant R01 DK089174; A. Rawlings, NIH/NHLBI grant T32 HL007024.
Expert Testimony: None declared.
Patents: None declared.
Role of Sponsor: The funding organizations played no role in the design of study, choice of enrolled patients, review and interpretation of data, or preparation or approval of manuscript.
- Received for publication June 27, 2014.
- Accepted for publication August 13, 2014.
- © 2014 American Association for Clinical Chemistry
References
In this issue
