Serum 1,5-anhydroglucitol when used with fasting plasma glucose improves the efficiency of diabetes screening in a Chinese population

Serum 1,5-anhydroglucitol (1,5-AG) levels can not only accurately reflect the mean blood glucose over the previous 1–2 weeks in diabetic patients but also offers the advantage of representing postprandial glucose. To evaluate the clinical significance of 1,5-AG in diabetes detection, especially when used in combination with fasting plasma glucose (FPG), a total of 3098 participants at high risk for diabetes (1467 men, 1631 women) were enrolled. A total of 1471 (47.5%) participants were diagnosed with diabetes, and the mean 1,5-AG level in the diabetic group was significantly lower than that in non-diabetic group [12.5 (7.8–17.5) μg/mL vs. 20.5 (15.3–26.4) μg/mL, P < 0.001]. The optimal cut-off point was 15.9 μg/mL, for which the sensitivity, specificity, and area under the curve (AUC) were 69.2%, 72.3%, and 0.781, respectively. For the combination of FPG and 1,5-AG, the sensitivity, specificity, and AUC improved to 82.5%, 83.5%, and 0.912, respectively. This method helped 75.8% of the participants avoid an oral glucose tolerance test (OGTT), reducing the need to carry out the OGTT by 43.9% compared to the use of the FPG criterion only. In conclusion, the addition of FPG to serum 1,5-AG improves the efficiency of diabetes screening in the Chinese population.


Results
Clinical characteristics of study participants. The characteristics of the 3098 subjects enrolled in the present study, including 1471 (47.5%) with diabetes and 1627 (52.5%) without diabetes, are shown in Table 1. The numbers of participants with isolated postprandial hyperglycemia, isolated elevated glycated hemoglobin A 1c (HbA 1c ), and elevated 2-hour postload plasma glucose (2hPG) and HbA 1c along with a FPG <7.0 mmol/L were 438 (29.8%), 111 (7.5%), and 227 (15.4%), respectively. These results indicated that 52.8% (776/1471) of participants with diabetes would be missed in screening based on merely FPG detection. Age, systolic blood pressure (SBP), body mass index (BMI), HbA 1c , FPG, and 2hPG levels were significantly higher in patients with diabetes than in participants without diabetes in both genders (all P < 0.05). In addition, in female participants, the diastolic blood pressure (DBP) was significantly higher in those with diabetes than in those without diabetes (P < 0.05).

Discussion
Our study indicated that the combined use of FPG and serum 1,5-AG criteria can improve the diagnostic efficiency of diabetes screening in a Chinese population. With this approach, OGTT could be avoided in 75.8% of the participants, and thus, the need for OGTT was reduced by 43.9% compared to that when only the FPG criterion was used. Also, the optimal cut-off point for serum 1,5-AG in diabetes screening was 15.9 μg/mL. For the combination of FPG with serum 1,5-AG, the sensitivity, specificity, and AUC were 82.5%, 83.5%, and 0.912, respectively. As for the combination of FPG and HbA 1c , the sensitivity, specificity, and AUC were 74.4%, 91.2%, and 0.911, respectively. There was no significant difference in the AUCs (P > 0.05). However, the specificity and sensitivity differed significantly from each other (all P < 0.001). Abbreviation: AUC: area under the curve; FDR: first degree relative with diabetes; FPG: fasting plasma glucose; HbA 1c : glycated hemoglobin A 1c ; ROC: receiver operating characteristic.
Scientific RepoRtS | 7: 11968 | DOI:10.1038/s41598-017-12210-z Then, we further compare the predictive value of the combination of FPG and 1,5-AG with that of FPG and HbA 1c . When compared with the combination of FPG and HbA 1c , there's no significant difference between the AUC of these two combined criteria. However, the sensitivity of the combination of FPG and 1,5-AG was higher than that of FPG and HbA 1c , and the specificity was lower. As a screening method, the ability to predict the positive population (i.e. sensitivity) is more important than specificity when a certain screening efficiency is ensured. Thus, the combination of FPG and 1,5-AG is suitable for diabetic screening. Screening strategies for detecting diabetes by an OGTT after use of (a) the FPG criteria, (b) the combination of FPG and HbA 1c , or (c) the combination of FPG and 1,5-AG. The proportions of the study population in specific diagnostic categories were: (a) FPG < 5.6 mmol/L to exclude and FPG ≥ 7.0 mmol/L to diagnose diabetes, with 43.1% participants needing an OGTT to confirm the diagnose. The sensitivity, specificity, PPV, and NPV for this strategy were 47.3%, 100%, 100%, and 67.7%, respectively. (b) As for the combined criteria of FPG and HbA 1c criteria, FPG < 5.6 mmol/L and HbA 1c < 5.7% to exclude diabetes, and FPG ≥ 7.0 mmol/L and/or HbA 1c ≥ 6.5% to confirm diabetes, 46.8% participants needed an OGTT. The sensitivity, specificity, PPV, and NPV for this strategy were 70.2%, 100%, 100%, and 78.8%, respectively. (c) With the combined FPG and 1,5-AG criteria, which are FPG < 5.6 mmol/L and 1,5-AG > 15.9 μg/mL for exclusion of diabetes, and FPG ≥ 7.0 mmol/L and/or 1,5-AG ≤ 15.9 μg/mL for the diagnosis of diabetes, only 24.2% participants needed an OGTT. The sensitivity, specificity, PPV, and NPV for this strategy were 78.7%, 72.3%, 72.0%, and 78.9%, respectively. Abbreviations: DM: diabetes mellitus; FPG: fasting plasma glucose; HbA 1c : glycated hemoglobin A 1c ; NPV: negative predictive value; OGTT: oral glucose tolerance test; PPV: positive predictive value.
The level of 1,5-AG, a polyol in the blood, is stable in individuals, with little fluctuation and no diurnal difference. 1,5-AG is mainly reabsorbed through the renal tube and maintained at a high level in normal circumstances. When the blood glucose exceeds the renal glucose threshold, the absorption of 1,5-AG will be inhibited by glucose [13][14][15] . Yamanouchi et al. 16 found that the plasma 1,5-AG level multiplied by the urinary glucose level was relatively constant, indicating that 1,5-AG is negatively correlated with blood glucose. Therefore, serum 1,5-AG measurement can both provide information about glycemic control and reflect short-term (over 1-2 weeks) glucose fluctuations 10,11 . In addition, another study conducted by Kim et al. 17 showed that serum 1,5-AG might be a reliable glycemic index in mild-to-moderate renal dysfunction as well. Our study also found that the serum 1,5-AG level was significantly negatively correlated with FPG, 2hPG, and HbA 1c levels.
Previous studies have shown that glycemic distributions vary among different races. For example, in the European population, diabetes patients mainly present with increased FPG 7,8 , whereas in the Asian population, diabetes patients exhibit postprandial hyperglycemia 18 , which was also observed in the current study. It was reported that 1,5-AG is clearly superior to HbA 1c in the accessment of postprandial glucose levels, especially for patients with HbA 1c < 8% 17,18 . Similarly, the current study also found that the serum 1,5-AG levels correlated negatively with glucose levels, and the association with 2hPG was superior to the association of FPG with serum 1,5-AG. In addition, serum 1,5-AG was more sensitive for diabetes in cases with 2hPG ≥ 9.0 mmol/L or HbA 1c < 8.5%.
The current study, for the first time, indicates the combination of FPG and serum 1,5-AG can guide diabetes screening. Based on FPG criteria only, 43.1% of the study participants would require an OGTT for definitive diagnosis, whereas with the combined use of FPG with serum 1,5-AG for detecting diabetes, only 24.2% of the study participants would need an OGTT. These findings indicate that the number of subjects requiring an OGTT can be reduced by more than 40% based on the combined use of FPG and 1,5-AG criteria, making screening for diabetes more sensitive and convenient. Our results also showed that the optimal cut-off point for 1,5-AG for the detection of diabetes in our Chinese population was 15.9 μg/mL. Shirasaya et al. 19 conducted a study in 891 Japanese subjects undergoing routine health examination and reported that the optimal cut-off point for 1,5-AG was 17.1 μg/ mL. Another related study in Japan (with 77 enrolled outpatients) showed that a serum 1,5-AG concentration of 14.2 μg/mL would be the best cut-off point for diabetes detection 20 . Based on the results outlined above, the ideal cut-off point for serum 1,5-AG in diabetes screening in Asian populations might fall in the range of 14.2-17.1 μg/ mL, and specific cut-off points for specific populations require further research. Interestingly, a study conducted by Wang et al. 21 based on a community population in Nanjing (25.5% of the participants were diagnosed with diabetes), the optimal cut-off point for serum 1,5-AG was 11.18 μg/mL. As mentioned in the discussion part, the gender imbalance and small sample size may explain this discrepancy. Moreover, we further performed a subgroup analysis based on FDR and found no significant difference in the screening efficacy of 1,5-AG between individuals with FDR and those without FDR.
There are some limitations in the current study. First, this was a single-center study, and the results need to be confirmed in a larger multi-center study. Second, further prospective research is needed given the cross-sectional design of this study. In addition, the present study population was selected from outpatients at high risk of diabetes, and future research is needed to confirm the findings in a larger cohort of general population.
In conclusion, we confirmed that the serum 1,5-AG level offers the advantage of reflecting postprandial glucose. This might explain why the combination of FPG and serum 1,5-AG criteria improved the early detection of diabetes in our Chinese population. In addition, the use of both FPG and serum 1,5-AG criteria helps to determine whether an OGTT is needed, allowing 75.8% of the participants to avoid an OGTT.

Materials and Methods
Study population. A total of 3098 individuals (1467 men, 1631 women; age range 20-80 years) who underwent a 75-g OGTT in the outpatient clinic of the Department of Endocrinology and Metabolism of Shanghai Jiao Tong University Affiliated Sixth People's Hospital between January 2011 and March 2017 were enrolled in the study. All participants had no history of diabetes mellitus or impaired glucose regulation, no current use of hypoglycemic agents (α-glucosidase inhibitors, canagliflozin), and no history of use of some traditional Chinese medicines such as polygala tenuifolia and senega syrup therapy. In addition, those with chronic renal failure, chronic liver dysfunction, and acute infection were also excluded from the study. This study was approved by the Ethics Committee of Shanghai Jiao Tong University Affiliated Sixth People's Hospital, and was in accordance with the 1964 Helsinki declaration and its later amendments or comparable ethical standards. Informed consent was obtained from all individual participants included in the study.
Anthropometric and biochemical assessments. All participants received a comprehensive physical examination, including height, body weight, and blood pressure. BMI was calculated as weight/height 2 (kg/m 2 ). Blood samples were obtained from all participants in the morning after a 10-h overnight fast to measure the levels of FPG, HbA 1c and 1,5-AG. A 75-g OGTT was administered to each patient to assay the 2hPG. Standard laboratory measurements were performed 22 . Serum 1,5-AG levels were measured by an enzymatic method (GlycoMark; GlycoMark Inc., New York, NY, USA) on a 7600 autoanalyzer (Hitachi, Tokyo, Japan) with inter-and intra-assay coefficients of variation (CVs) of <3.5% and <2.5%, respectively.
Definition and diagnostic criteria. According to the 2010 ADA criteria 23 , those who meet the diagnostic criteria of FPG ≥ 7.0 mmol/L, 2hPG ≥11.1 mmol/L, or HbA 1c ≥ 6.5% should be diagnosed with diabetes. FDR indicates the participant had at least one first-degree relative (parent, child, or sibling) with diabetes 24 .
Scientific RepoRtS | 7: 11968 | DOI:10.1038/s41598-017-12210-z Statistical analysis. SPSS version 19.0 (SPSS, Inc., Chicago, IL, USA) and MedCalc version 15.2 were used for statistical analysis. Based on a normality test, normally distributed data are presented as mean ± standard deviation values, and skewed data are presented as median with interquartile ranges. Categorical variables are presented as percentages (%). Unpaired Student's t test and Chi-squared test were carried out for inter-group comparisons of normally distributed data, and the Wilcoxon rank sum test along with the Kruskal-Wallis test were used for inter-group comparisons of non-normally distributed variables. Spearman correlation analysis and multiple stepwise regression analysis were conducted to identify independent factors. The ROC curve was generated to analyze the value of serum 1,5-AG in screening diabetes. The optimal cut-off point was confirmed based on the Youden index. A two-tailed P value < 0.05 was considered to be statistically significant.
Ethical approval. This study was approved by the Ethics Committee of Shanghai Jiao Tong University Affiliated Sixth People's Hospital, and was in accordance with the 1964 Helsinki declaration and its later amendments or comparable ethical standards.
Informed consent. Informed consent was obtained from all individual participants included in the study.