Postprandial Glucose Levels Are Better Associated with the Risk Factors for Diabetes Compared to Fasting Glucose and Glycosylated Hemoglobin (HbA1c) Levels in Elderly Prediabetics: Beneficial Effects of Polyherbal Supplements—A Randomized, Double-Blind, Placebo Controlled Trial

Backgrounds Prediabetes is a condition in which a person's blood glucose levels are higher than normal physiological levels but lower compared to patients with diabetes. Up to 70% of individuals with prediabetes will eventually develop diabetes. To date, there have been no pharmaceutical drugs to treat diabetes. It is believed that early diagnosis and nonpharmacological intervention for prediabetes are critical for effective prevention of diabetes. Most individuals with prediabetes remain undiagnosed even after being evaluated using the standard tests for fasting glucose (FG) and HbA1c. We investigated if postprandial glucose levels (2h-PG) were associated with pre/diabetes and if polyherbal supplements could be beneficial for individuals with prediabetes. Materials and Methods 100 elderly individuals with impaired 2h-PG or fasting glucose levels were recruited to receive either a 12-week supplement of GlucoVita (an antioxidative polyherbal formulation) (n=50) or placebo (n=50). Results No baseline differences were observed for FG, HbA1c, or 2h-PG. Individuals who received a twelve-week administration of GlucoVita supplements had significantly reduced 2h-PG (8.15±1.67 versus 7.35±2.06 mmol/l, P<0.05) levels compared to individuals in the placebo group. In addition, HbA1c levels were lower in individuals who received GlucoVita (5.81±0.49 %) compared to the individuals in the placebo group (6.00±0.51%) (P=0.08) after 12-weeks. Stratified analysis, based on impaired fasting glucose (IFG), 2h-PG, metabolic symptom, and age, demonstrated that, after the 12-week intervention, HbA1c levels were significantly lower in the GlucoVita administered group compared to the placebo group (IFG subgroup; 5.85±0.46%, n= 27 versus 6.14±0.50, n=33, P<0.05) and the metabolic symptom-free subgroup (5.73±0.45%, n=23 versus 6.04±0.52%, n=24, P<0.05). GlucoVita also reduced FG in individuals with normal 2h-PG (6.37±0.27 versus 6.08±0.38 mmol/l, P<0.05). Baseline 2h-PG levels, but not HbA1c or FG levels, were significantly correlated with body weight, waist circumference, and BMI (r=0.25, P<0.05; r=0.31, P<0.01; r=0.22, P<0.05, respectively). Conclusion 2h-PG levels were better associated with body weight, waist circumference, and BMI risk factors compared to FG and HbA1c levels in elderly individuals with prediabetes. Polyherbal formulation GlucoVita supplements improved 2h-PG and HbA1c levels only in elderly individuals who were overweight but were symptom-free and under 65 years of age. Due to the small cohort size of this pilot study, future studies are required to validate our findings.


Introduction
More than 330 million individuals worldwide are estimated to have prediabetes and up to 70% will develop diabetes within a decade of initial diagnosis [1]. According to the US Centers for Disease Control (CDC), 84 million Americans (more than 1 out of 3) have prediabetes. Of these, 90% are unaware of their prediabetic status and up to 70% of them will eventually progress to diabetes [2]. The average risk of developing diabetes is approximately 5-10% per year in individuals with IFG or IGT, which is 10-15 times higher compared to individuals with normoglycemia.
Prediabetes is a condition in which a person's blood glucose level is higher than normal physiological levels but lower than individuals with diabetes. Prediabetes increases the risk of diabetes and cerebrocardiovascular disorders, which are the major causes of death worldwide [3].
Despite the worldwide prevalence of prediabetes and diabetes, many patients remain undiagnosed. This may be due to the current reference standards for fasting glucose (FG), or the inconvenience for individuals taking the 2horal glucose tolerance test (OGTT). In 2010, the American Diabetes Association approved the more expensive assay of glycated hemoglobin A1c (HbA1c) as one of the three tests for diabetes and prediabetes screening. However, recent studies have shown substantial discordance between results from FG, OGTT, and HbA1c, with OGTT probably being the more accurate diagnostic test for pre/diabetes associated with age, BMI, physical activity, income, education, race/ethnicity, etc. [4][5][6]. Conversely, recent studies have shown that biomarkers for food intake and nutrient status were associated with glucose tolerance status and development of diabetes in the elderly [7]. The utility of these new dietary biomarkers for determining pre/diabetes status has not been comprehensively evaluated in different ethnic populations.
Oxidative stress, inflammation, and peripheral insulin resistance lead to pancreatic beta cell overproduction; however eventual depletion of insulin secretion is thought to underlie the development of prediabetes and diabetes [8][9][10][11][12][13]. Overweight, being 45 years or older, and lower physically active are associated with type 2 diabetes and are also the main risk factors for prediabetes [2,14]. Modification in lifestyle, increase in physical activity, having a healthy diet, and maintaining a healthy body weight can lower the risk of developing prediabetes and diabetes [15][16][17]. However, many people have difficulties to adhere to these recommendations.
Herbal supplements have long been used for disease prevention and treatment. Recent cellular and animal model studies have suggested that herbal supplements could improve glucose metabolism and protect from oxidative stress-induced diabetes [18][19][20][21][22][23][24]. However, several of these findings were deduced from animal studies and have not been replicated and validated in human studies. So far, only a few clinical studies have been performed to validate the claims of the effectiveness of herbal ingredients on prediabetes. Numerous studies have reported the antidiabetic and antioxidative activities of momordica charantia (bitter melon) extracts in diabetic animal models [25][26][27][28][29][30][31][32]; however, other studies have reported insufficient evidence for the effects of momordica charantia for type 2 diabetes mellitus [33][34][35].
A recent study showed that daily intake of 1g of cinnamon for 12 weeks reduced fasting blood glucose and HbA1c levels in middle-aged type 2 diabetic patients [36]. The antidiabetic effects of cinnamon were reported to increase glucose uptake, improve insulin sensitivity in peripheral tissues, improve glycogen synthesis in the liver, restore pancreatic islets dysfunction, slow gastric emptying rates, and improve diabetic renal and brain disorders through multiple signaling pathways, including the PPARs, PI3K/IRS-1, GLUT4, and Nrf2 pathways [37].
In this study we compared the predictive value of FG, 2h-PG, and HbA1c levels in elderly Chinese individuals with prediabetes and evaluated the effects of 12-week administration of GlucoVita supplements, a polyherbal formulation, on these three blood parameters using a double-blinded, randomized placebo control trial design.

Materials and Methods
. . Study Participants. Patient recruitment was initiated on September 2, 2011, and completed on October 12, 2012, with the last follow-up observation completed on January 7, 2012. Individuals were recruited from the Sanlin Community Health Service Center, Pudong New District, Shanghai, China. All study procedures were conducted in accordance with the Helsinki Declaration of 1975 and were approved by the Shanghai Jiao Tong University School of Public Health Institutional Review Board. Written informed consent was obtained from all study participants prior to enrollment into the study. This study is part of a series investigations into the effects of dietary supplements in old adults with chronic health conditions. Similar research methodologies have been used and reported in previous studies [38].
Subjects who met the first and one of the following two criteria were eligible for study participation.

Inclusion Criteria
(1) Healthy males or females at least 50 years of age.
(2) Administered glucose lowing medication over the last 30 days.
. . Randomization and Blinding. Enrolled study participants were randomly assigned to the GlucoVita treatment Evidence-Based Complementary and Alternative Medicine 3 group or the placebo control group. The randomization was performed using a predetermined randomization code which was generated using a random number generator. Study participants and physicians were double-blinded for the treatment. Of the 100 enrolled participants, 88 of them completed the 12-week follow-up period, which included 43 in the GlucoVita group and 45 in the control group. Five individuals in the GlucoVita group and 4 in the placebo control group withdrew from the study due to objections from their family members. Two individuals in the GlucoVita group withdrew from the study due to diarrhea and nausea and one individual in the placebo group withdrew from the study due to respiratory symptoms (coughing).
Study participants in the control group received similarlooking capsules in color-coded bottles (white bottles for GlucoVita and yellow bottles for placebo control). Neither the study participants nor the physicians, including the study principal investigator, had knowledge of the specific color codes until the end of the study. Both the GlucoVita tablets and the placebo control tablets were manufactured and supplied by GardaVita5 Inc. (Costa Mesa, California, USA). Each study participant was instructed to take 1 tablet with meals, two times per day for 12 weeks, with a new batch of 4 Evidence-Based Complementary and Alternative Medicine  supplements dispensed every month during the follow-up sessions.

. . Demographic Information, Medical Examination, and
Patient Follow-Up. The date of birth, gender, status of impaired fasting glucose (IFG), impaired postprandial glucose tolerance (I2hPG), body weight index (BMI), status of hypertension, dyslipidemia, and the presence of diabetes-related symptoms (polydipsia, polyphagia, hyperuresis, underweight, overweight, and others, 1=less or none, 2= moderate, and 3= most severe) were recorded for each study participant at the time of the initial screening before randomization. Physical examinations that focused on 4 levels of general health conditions (better than average; average; declining health status; frequently sick, 1=true, 2=most true, 3=unsure, 4= most false, and 5= false) were evaluated for all study participants by a physician. A standard form was used to screen study participants and they were asked to return to the clinic on a weekly basis for the first month and on monthly basis for the remaining months. During followup, the physicians answered any concerns and evaluated compliance/adherence and recorded adverse events. Study participants were then resupplied with a new bottle of tablets.
. . Statistics Analysis. EpiData 3.1 software was used for data entry and SPSS 20 software was used for statistical analysis. Group data were presented as mean ± standard deviation or median ± Quartile Range (QR). Differences between the GlucoVita and placebo groups were compared using Student's t-test for quantitative variables with normal distribution or Chi-square for categorized variables. Appropriate stratification analysis was performed to control for confounding factors. Correlations between the variables were also determined for each of the subgroups. P<0.05 denoted statistical significance. All p-values reported were 2-sided.
. . HbA c Levels. No baseline differences in HbA1c levels were found between the GlucoVita and placebo groups (5.85±0.41 versus 5.92±0.42, P>0.05). However, after the intervention, HbA1c levels were lower in the GlucoVita group compared to the placebo group at a trend level (5.81±0.49 versus 6.0±0.51, P=0.08). Similarly, the group difference for the rate of change in HbA1c levels was at a trend level (0.05±0.31 versus -0.08±0.36, P=0.09).
Stratified analysis showed that, for the ≤65-year subgroup, HbA1c levels were slightly but significantly reduced in the GlucoVita subgroup after the 12-week intervention (5.73±0.43 versus 5.85±0.37, P<0.05) but were slightly increased in the placebo group (5.96±0.48 versus 5.89±0.39). HbA1c levels was lower at a trend level in the GlucoVita group compared to the placebo group for the ≤65-year subgroup (5.73±0.43 versus 5.96±0.48, P=0.06) after the 12week intervention (Table 4).
Similarly, within the metabolic symptom-free subgroups, HbA1c levels were significantly lower in the GlucoVita group compared to the placebo group after the 12-week intervention (5.73±0.45 versus 6.04±0.52, P<0.05) ( Table 4). Within the IFG subgroup, HbA1c levels were significantly lower in the GlucoVita group compared to the placebo group after the 12week intervention (5.85±0.46 versus 6.14±0.5, P<0.05). For the normal 2h-PG subgroup, HbA1c levels were lower in the GlucoVita group compared to the placebo group after the 12-week intervention (5.74±0.41 versus 5.99±0.36, P=0.08) ( Table 4).
The rate of reduction of HbA1c levels was greater in the GlucoVita group compared to the placebo group after the 12-week intervention at a trend level (0.05±0.31 versus -0.08±0.36 %, P<0.10). The group differences were significant for ≤65 years of age subgroup (0.12±0.28 versus -0.07±0.41%, P<0.05) ( Similarly, baseline HbA1c levels were significantly correlated with baseline fasting glucose levels for all participants (n=88, r=0.27, P<0.01) and in the placebo group (n=45, r=0.36, P<0.05), but only at a trend level in the GlucoVita group (n=43, r=0.19, P>0.05). No significant correlations were found between baseline HbA1c levels and baseline postprandial glucose levels (2h-PG) for all the groups.

Discussion
Prediabetes and type 2 diabetes have become a global epidemic, with people having prediabetes being undiagnosed and progressing towards diabetes. Hence there is an urgent need for early diagnosis and effective interventions to prevent and/or reverse the progression of prediabetes in to full blown diabetes. FG, OGTT, and HbA1c are the choice methods to determine abnormal glucose regulation. However, OGTT is too inconvenient of a method for many people and FG and HbA1c are not accurate predictors of prediabetes. Additionally, there are discordances between the results from these tests. Recent studies have shown that new biomarkers for food intake and nutrient status are associated with glucose tolerance status and the development of diabetes [7]. However, validation of any new biomarkers is a long and expensive process.
In this study, we found that postprandial glucose levels are closely associated with the risk factors for diabetes compared to FG and HbA1c levels in elderly individuals with prediabetes. Additionally, a significant reduction in 2h-PG levels and a trend reduction in HbA1c were found after 12-week administration of GlucoVita supplements. These beneficial effects of GlucoVita were primarily observed in individuals ≤ 65 years of age and in metabolic symptom-free individuals. GlucoVita supplements also significantly reduced FG levels in individuals with normal 2h-PG levels and reduced 2h-PG levels in individuals ≤ 65 years of age (but not in individuals >65 year). 2h-PG levels were also reduced in the normal FG subgroup but not in the impaired FG subgroup, suggesting that GlucoVita was more effective in individuals with early stage prediabetes.
Similarly, the beneficial effect of GlucoVita on HbA1c levels was significant only in individuals ≤ 65 years of age and in those without IGT and metabolic symptoms. These results indicate that polyherbal supplementation could be beneficial for the "relatively young" individuals during the early stages of prediabetes. Our results were in agreement with the recent study that indicated daily intake of cinnamon for 12 weeks reduced fasting blood glucose and HbA1c levels among middle-aged type 2 diabetic patients [36]. Although the underlying mechanisms remain to be elucidated, these antidiabetic effects of these herbal supplements may be through increasing glucose uptake and improving insulin sensitivity in peripheral tissues, as well as improving glycogen synthesis in the liver, restoring pancreatic islets dysfunction, slowing down gastric emptying rates, and improving diabetic renal and brain disorders, via multiple signaling pathways, including those of the PPARs, AMPK, PI3K/IRS-1, RBP4-GLUT4, and Nrf2 pathways [37].
In this study, we also found that baseline and postintervention HbA1c levels were significantly correlated with FG levels (r=0.28, P<0.01; r=0.55, P<0.01, respectively) rather than with 2h-PG levels (r=0.14, P>0.05; r=0.37, P<0.01, respectively) or other risk factors in elderly individuals with prediabetes. This supports the notion that HbA1c levels are a good indicator of FG. HbA1c is an index of average glucose over the preceding weeks-to-months and is a more stable indicator of glucose status rather than fasting glucose or postprandial glucose. This is because the erythrocyte (red blood cell) lifespan averages about 120 days. The HbA1c levels at any time point are a result of all circulating erythrocytes, from the oldest (120 days old) to the newly generated. Because HbA1c is a weighted average of blood glucose levels of the preceding 120 days (the lifespan of erythrocytes), newly generated erythrocytes contribute more to determining current glucose levels in the immediate preceding days compared to erythrocytes generated 90-120 days earlier. Thus, HbA1c levels could increase or decrease relatively quickly following a clinically significant change in acute glucose levels.
The known risk factors of prediabetes such as body weight, waist circumference, and BMI are significantly correlated with baseline and postintervention 2h-PG levels rather than with FG or HbA1c. Our results are in-line and suggest that 2h-PG rather than FG or HbA1c should be used as the more reliable indicator to screen elderly individuals with prediabetes. HbA1c levels became highly correlated with FG levels (r=0.53, P<0.01) and 2h-PG levels (r=0.31, P<0.01) after 12-week administration with GlucoVita, supporting a beneficial role of GlucoVita supplements for elderly individuals.
Limitations. There were several limitations of this pilot study and our results should be interpreted with caution. The number of study participants was small and the observational follow-up period was short to comprehensively demonstrate that 2h-PG levels was a better indicator compared to FG and HbA1c levels in predicting the development of diabetes for the different subgroups (age, body weight, symptoms, and BMI) of elderly individuals with prediabetes. Further studies are required to investigate if 2h-PG is more sensitive and reliable compared to FG/HbA1c levels during the prediabetic stage to predict the development of diabetes. In addition, further studies should be performed to determine if herbal supplements could be effective to improve 2h-PG levels at the preclinical stage of diabetes.

Conclusions
In this study, 2h-PG but not FG and HbA1c levels were significantly correlated with body weight, waist circumference, and BMI (the known risk factors of diabetes in elderly individuals with prediabetes). Twelve-week administration of GlucoVita supplements significantly improved 2h-PG levels only in prediabetics who were symptom-free and were under 65 years .595 * * .431 * * 1 * , Pearson correlation was significant at the 0.05 level (2-tailed). * * , Pearson correlation was significant at the 0.01 level (2-tailed). of age. As most prediabetics would progress to diabetes in less than 10 years, our study demonstrated a potential benefit of herbal supplements on 2h-PG levels for the management of preclinical stage of diabetes.