Glucose Fluctuations in Association With Oxidative Stress Among Children With T1DM: Comparison of Different Phases

Transcription

1 ORIGINAL ARTICLE Glucose Fluctuations in Association With Oxidative Stress Among Children With T1DM: Comparison of Different Phases Xi Meng,* Chunxiu Gong,* Bingyan Cao, Xiaoxia Peng, Di Wu, Yi Gu, Liya Wei, Xuejun Liang, Min Liu, Wenjing Li, and Chang Su Department of Endocrinology, Genetics, and Metabolism (X.M., C.G., B.C., D.W., Y.G., L.W., X.L., M.L., W.L., C.S.), Ministry of Education National Key Discipline of Pediatrics (X.M., C.G., B.C., D.W., Y.G., L.W., X.L., M.L., W.L., C.S.), and Center of Clinical Epidemiology and Evidenced Based Medicine (X.P.), Beijing Children s Hospital, Capital Medical University, Beijing , P.R. China Context: Previous studies on the association between glucose fluctuations and oxidative stress in diabetes mellitus were all conducted in adults, and most study participants had type 2 diabetes mellitus. Objective: The study objective was to examine the association of glucose fluctuation and oxidative stress in children with type 1 diabetes mellitus (T1DM) across different phases. Design: This was a case-control study. Setting: The setting was the Beijing Children s Hospital ( ). Patients: Children treated for T1DM were divided into three study groups, including group A (newly diagnosed cases in the acute metabolic disturbance phase), group B (the honeymoon phase), and group C (the long-standing phase). Healthy control children were matched to the T1DM patients by age and sex. Interventions: There were no interventions. Main Outcome Measures: The 24-hour urinary free 8-iso-prostaglandin F2 to creatinine (8-iso- PGF2 /Cr) ratio indicated oxidative stress. Glucose fluctuation parameters (GFPs) included mean blood glucose levels, standard deviation of daily blood glucose levels, mean amplitude of glucose excursions, and incremental area under the curve for postprandial glucose. Results: In each study group, 8-iso-PGF2 /Cr and all GFPs in children with T1DM were significantly higher than those in normal controls. 8-iso-PGF2 /Cr was significantly correlated with all GFPs in all three T1DM groups. A stronger association with 8-iso-PGF2 /Cr for mean amplitude of glucose excursions than for glycated hemoglobin was observed in both the acute metabolic disturbance and long-standing phases of T1DM. Conclusions: Glucose fluctuations were positively associated with oxidative stress in T1DM patients in different phases. Glucose fluctuations may have a stronger effect than sustained chronic hyperglycemia on triggering oxidative stress, but coexisting high levels of blood glucose may be required. (J Clin Endocrinol Metab 100: , 2015) ISSN Print X ISSN Online Printed in U.S.A. Copyright 2015 by the Endocrine Society Received July 9, Accepted February 20, First Published Online February 25, 2015 * X.M. and C.G. contributed equally to this work. Abbreviations: BMI, body mass index; CGM, continuous glucose monitoring; CGMS, CGM system; Cr, creatinine; DBP, diastolic blood pressure; HbA1c, glycated hemoglobin; IAUC, incremental area under the curve; 8-iso-PGF2, 8-iso-prostaglandin F2 ; MAGE, mean amplitude of glucose excursions; MBG, mean blood glucose; SBP, systolic blood pressure; SDBG, standard deviation of blood glucose; T1DM, type 1 diabetes mellitus; T2DM, type 2 diabetes mellitus jcem.endojournals.org J Clin Endocrinol Metab, May 2015, 100(5): doi: /jc

2 doi: /jc jcem.endojournals.org 1829 Chronic complications of diabetes, including microvascular and macrovascular changes, are the leading cause of diabetes-related death and disability. Preventing or delaying the development of these complications is very important. Recent evidence shows that glucose fluctuations may influence the development of diabetic complications (1, 2), possibly by triggering oxidative stress (3), although the data from the HEART2D trial and the Verona Diabetes Study did not seem to support this (4, 5). An association of glucose fluctuation and oxidative stress has been observed in both in vitro cell culture and animal studies (6, 7). Monnier et al (8) reported an effect of glucose fluctuations on oxidative stress in patients with noninsulin-treated type 2 diabetes mellitus (T2DM), but this was not confirmed in patients with type 1 diabetes mellitus (T1DM) (9). Monnier et al (10) suggested that these conflicting findings might be explained by an inhibitory effect of insulin therapy on oxidative stress. However, a subsequent study on insulin therapy and oxidative stress did not support that conclusion but found that the relationship was not a simple one (11). There are few clinical studies on the effects of glucose fluctuations in patients with T1DM, either in China or other countries. Therefore, further research in this area is needed (12, 13). Previous studies on the association between glucose fluctuations and oxidative stress in diabetes mellitus were all conducted in adults, and most study participants had T2DM. The findings obtained in adults with T2DM are likely to be influenced by smoking status (14), hypertension (15), hyperlipidemia (16), obesity (17), and other variables that may affect the level of oxidative stress. The influence of confounding variables increases the possibility of experimental bias and makes it difficult to interpret the results in adult studies with small sample sizes. In this study, we included only children with T1DM who were free of smoking, hypertension, hyperlipidemia, and complications of diabetes or obesity in the study population. This case-control study aimed to analyze the patterns of glucose fluctuations in children with T1DM across different phases (ie, the acute metabolic disturbance phase, the honeymoon phase, and the long-standing phase). The associations of several glucose fluctuation parameters with oxidative stress were investigated. Subjects and Methods Subjects and study design The study protocol was approved by the Institutional Review Board of the Beijing Children s Hospital. Signed informed consent was obtained from each participant or their guardian after being told of the research aims and study methods. The study was conducted from 2010 to 2013 and enrolled 60 T1DM patients from 5 to 18 years of age. T1DM was diagnosed according to the International Society for Pediatric and Adolescent Diabetes Clinical Practice Consensus Guidelines (18). Participants were stratified to three groups by T1DM phase: newly diagnosed cases in the acute metabolic disturbance phase (group A; n 20), cases in the honeymoon phase (group B; n 20), and cases in the long-standing phase (ie, more than 5 years since diagnosis; group C; n 20). Group A children were enrolled 1 week after diabetic ketoacidosis had been rectified or insulin therapy had begun. The inclusion criteria for group B (19) are: 1) children requiring insulin treatment at the dose of 0.5 IU/kg d in combination with glycated hemoglobin (HbA1c) 7.5%; or 2) children having used an even lower limit for insulin requirement 0.5 IU/kg d. Children with acute or chronic complications of diabetes were excluded. The control group was recruited from children with idiopathic short stature, frequency-matched to the T1DM patients by age and sex, and with normal glucose metabolism (fasting blood glucose 5.6 mmol/l [100 mg/dl] and 2-h postprandial blood glucose 7.8 mmol/l [140 mg/dl]) in the Department of Endocrinology, Beijing Children s Hospital, during the same period. Participants were excluded if they had: 1) taken medications that could affect glucose metabolism (eg, corticosteroids, hypoglycemic agents, diuretics, antipsychotics) or vitamin C and vitamin E supplements within 1 month before enrollment; 2) severe liver or kidney dysfunction, heart failure, severe trauma, infections; or 3) other diseases that could affect glucose metabolism (eg, Cushing s syndrome, hyperthyroidism). Study procedures All T1DM patients and controls underwent continuous glucose monitoring for 2 days using the MiniMed continuous glucose monitoring system (CGMS; Medtronic). The monitor was implanted by trained nurses. Fingertip blood glucose levels were measured (OneTouch Ultra glucose meter; Johnson & Johnson) eight times, and the values were used to titrate the CGMS meters. The CGMS monitor was removed after 3 days, and the data that had been collected were downloaded and analyzed using CGMS Software version 3.0. Children with T1DM did not receive any insulin dose adjustments while wearing the CGMS monitor. All of the participants were provided diet and exercise management by dieticians. The total daily calorie needs of children in group A were determined as: age (y) kcal. For children in groups B and C, the calorie requirements were based on a dietician s assessment of nutritional status before hospital admission and were served in measured quantities of food at fixed mealtimes. Children were instructed to engage in low-intensity physical activity 20 minutes after meals and before bedtime (eg, walking in the ward for 15 minutes). Twenty-four-hour urine samples were collected for 2 consecutive days during the continuous glucose monitoring. We pooled the 24-hour urine samples, recorded the total volume, mixed the sample thoroughly, and retained 10 ml for storage at 80 C until laboratory testing. Fasting venous blood samples were collected on the first day. All participants were given standardized physical examinations that included measurement of height, weight, and systolic and diastolic blood pressure, and information about disease history and insulin dose was collected.

3 1830 Meng et al Glucose Fluctuations and Oxidative Stress in T1DM J Clin Endocrinol Metab, May 2015, 100(5): Measurement of glucose fluctuations Mean blood glucose (MBG) MBG was the average of 24-hour blood glucose levels obtained from CGMS monitoring. Daily glycemic variability Twenty-four-hour standard deviation of blood glucose (SDBG) reflects the extent of deviation from the MBG level and illustrates the dynamic dispersion of daily blood glucose levels. Twenty-four-hour mean amplitude of glycemic excursions (MAGE) provides the mean value of all the amplitudes of glycemic excursions that exceed one SDBG. It can make up for the SDGB shortcoming of not distinguishing between major and minor glycemic excursions. Postprandial glucose excursion The size of the incremental area under the curve (IAUC) was used as an indicator of postprandial glucose characteristics and postprandial glucose control. Instantaneous blood glucose levels determined before each of the three daily meals were set as baselines, the 3-hour postprandial glucose levels after each meal were plotted, and mean values of the areas under the glucose curves were calculated. Measurement of oxidative stress 8-Iso-prostaglandin F2 (8-iso-PGF2 ), an abundant and stable iso-prostaglandin, is a reliable indicator of lipid oxidative stress (20, 21). The ratio of 24-hour urinary free 8-iso-PGF2 to creatinine (Cr) (8-iso-PGF2 /Cr, or 8-iso-PGF2 excretion rate) was used here to measure differences in oxidative stress. Urinary free 8-iso-PGF2 was measured by an ELISA (Cayman Chemical Co., Ann Arbor, MI) using an automated, multifunctional microplate reader (Labsystems Dragon Multiskan MK3). The intra- and interassay coefficients of variation were both 5%, and the minimum detectable concentration was 2.7 pg/ml. Urinary Cr was measured with a two-point endpoint method using an automated biochemical analyzer (Toshiba 120; reagents, Prodia Diagnostics). The relative deviation for detection accuracy was within 10%. The intra- and interassay coefficients of variation were both 5%. Other laboratory tests Fasting blood samples were tested by a central laboratory in our hospital. The following assays were carried out: fasting plasma glucose (hexokinase method, Olympus AU2700 automatic biochemical analyzer), C-peptide (solid-phase, competitive chemiluminescent enzyme immunoassay, Immulite Analyzer), HbA1c (ion exchange HPLC, SYSMEX automated HbA1c analyzer G7), liver function (alanine aminotransferase, Dade Dimension method), renal function (blood urea nitrogen, phthalaldehyde colorimetry; Cr, enzymatic assay), and lipid profiles (triglycerides, enzyme assay; total cholesterol, enzyme assay; high-density lipoprotein-cholesterol, the direct method; low-density lipoprotein-cholesterol, Friedewald formula). Liver and kidney function tests and lipid profiles were done with an Olympus AU2700 automatic biochemical analyzer. Statistical analysis Numerical data were expressed as means SD. Group means of continuous variables were compared by the t test or Wilcoxon test. ANOVA was used for comparisons of multiple group means of continuous variables. Analysis of covariance or the Kruskal Wallis rank sum test and the Q test were used for comparisons between two groups among multiple groups. Categorical data were expressed as percentages, and differences in the distribution of categorical data were tested by the 2 test. Multiple correlation and multiple linear regression were used to determine the significance of associations among the study variables. Multicollinearity diagnosis was also performed for all regression models. P.05 was considered statistically significant. All statistical analyses were carried out using the SPSS version 15.0 software package (SPSS Inc). Results Participant characteristics The characteristics of the T1DM patients (60 cases) and the case-control participants (20 cases) are shown in Table 1. Body mass index (BMI), systolic blood pressure (SBP), diastolic blood pressure (DBP), fasting blood glucose, HbA1c, C-peptide, and lipid profiles were within their normal ranges in the control group. None of the participants with T1DM were obese or had hypertension or dyslipidemia. The controls were age-matched to the T1DM patients (P.596). There were no significant differences in age among the control group, group A, and group B, but the mean age of group C was significantly older than that of the other three groups of children. In this study, pubertal children are all from the honeymoon phase and the long-standing phase. In the honeymoon phase, three cases can be classified as Tanner II. In the long-standing phase, one case is equivalent to Tanner II, two to Tanner III, four to Tanner IV, and one to Tanner V. No significant differences were observed in gender distribution, SBP, or DBP across all groups. There were no significant differences in the BMI of the normal control group compared with group A or group B, or of group B compared with group C. However, BMI was significantly greater in group C and group B than in group A. The BMI was also significantly higher in group C than in the control group; the difference remained significant after adjusting for age. Patients in group B all received two injections daily of a mixture of short- and intermediate-acting insulin (before breakfast and dinner). Those in group A were treated with an insulin pump regimen. Forty percent of group C patients used a regimen of shortand intermediate-acting insulin mixture, whereas the remaining patients used a regimen of basal-bolus (one-third of pump, two-thirds of intermediate-acting/long-acting insulin/ basal analog once or twice daily, and rapid-acting or regular boluses with meals and snacks).

4 doi: /jc jcem.endojournals.org 1831 Table 1. Characteristics of Children With T1DM at Different Phases and of the Control Group Variable Control Group Group A (Acute Metabolic Disturbance Phase) Glucose fluctuation parameters Comparisons of MBG, daily blood glycemic excursion, postprandial blood glycemic excursion, and oxidative stress levels among the T1DM patient groups and the control participants are shown in Table hour MBG The 24-hour MBG in the control group ( mmol/l, or mg/dl) was significantly lower than that in the T1DM patients across all groups (P.01 vs group A and group C; P.001 vs group B). The MBG of group C was significantly higher than that of group A (P.009) and group B (P.005). But MBG of group A Group B (Honeymoon Phase) Group C (Long-Standing Phase) F P No. of subjects Disease duration, y ND b b,c Age, y a,b,c Gender, male/female, n 13/7 12/ / BMI, kg/cm b a,b SBP, mm Hg DBP, mm Hg FBG, mmol/l a a a,b,c HbA1c, % a a,b a,b,c C-peptide, ng/ml ND b b,c TG, mmol/l CHO, mmol/l LDL-C, mmol/l HDL-C, mmol/l a a a Insulin treatment dose, IU/kg/d b b,c Abbreviations: F, result of variable comparisons among groups; FBG, Fasting blood glucose; TG, triglycerides; CHO, total cholesterol; LDL-C, lowdensity lipoprotein cholesterol; HDL-C, high-density lipoprotein cholesterol; ND, not determined. Data (except for gender) are expressed as means SD. Unit conversions: to convert total, HDL, and LDL cholesterol to mg/dl, multiply by 38.67; to convert triglycerides to mg/dl, multiply by 88.54; to convert glucose to mg/dl, multiply by a P.05 vs control group. b P.05 vs group A. c P.05 vs group B. was not very significantly higher than that of group B (P.811). Parameters of glycemic excursion All three T1DM groups had significantly higher SDBG, MAGE, and IAUC values than the control group. SDBG and MAGE levels were highest in group C and were significantly higher than in groups A and B (P.05). No significant differences in SDBG and MAGE levels were observed between groups A and B. The IAUC level was highest in group B and was significantly higher than that in group C (P.05); differences in the IAUC levels observed in groups B and Table 2. Glucose Fluctuation Parameters and 8-Iso-PGF2 /Cr in Children With T1DM at Different Phases and in the Control Group Variables Control Group Group A (Acute Metabolic Disturbance Phase) Group B (Honeymoon Phase) Group C (Long-standing Phase) F P MBG, mmol/l a a a,b,c SDBG, mmol/l a a a,b,c MAGE, mmol/l a a a,b,c IAUC, mmol/l h a a a,c h 8-iso-PGF2 /Cr, pg/mg a a a,b,c Data are expressed as means SD. To convert glucose to mg/dl, multiply by F, result of variable comparisons among groups. a P.05 vs control group. b P.05 vs group A. c P.05 vs group B.

5 1832 Meng et al Glucose Fluctuations and Oxidative Stress in T1DM J Clin Endocrinol Metab, May 2015, 100(5): Figure 1. Twenty-four-hour continuous glucose concentrations and SD curves in children with different phases of T1DM and in normal controls. C were not significantly different than those seen in group A. 24-hour blood glucose and SD curves The glycemic excursion curves (Figure 1) show that group C had the largest SD, suggesting a large degree of variation in daily blood glucose levels. The 24- hour blood glucose and SD curves of the control group were significantly lower than those of all three T1DM groups. The SD curves of groups A and B were similar. Oxidative stress: 8-iso-PGF2 excretion rate The 8-iso-PGF2 /Cr was significantly higher in all T1DM groups than in the control group. The urinary 8-iso-PGF2 /Cr ratios in groups A, B, and C were 3.0, 2.32, and 5.19 times larger, respectively, than that of the control group. Group B had the lowest 8-iso- PGF2 /Cr value; group C had the highest (P.01, groups A and B). These differences remained significant after adjusting for age, sex, BMI, blood pressure, and blood lipids. Differences in 8-iso-PGF2 /Cr between groups A and B were not significant. Correlation and regression analysis of glucose monitoring parameters and oxidative stress Correlation analysis In all three T1DM groups, there was a significant positive correlation between MBG, SDBG, MAGE, and IAUC with 8-iso-PGF2 /Cr. Figure 2 shows the correlations between 8-iso-PGF2 /Cr and blood glucose fluctuation parameters in each of the three T1DM groups; when the three groups were combined, the correlation coefficients were for MBG, for SDBG, for MAGE, and for IAUC (all P.001). In the control group, no significant correlations were found between MBG, SDBG, MAGE, or IAUC and 8-iso-PGF2 /Cr. Multiple linear regression analysis Multiple linear regression analysis revealed that disease duration, age, HbA1c, MAGE, and IAUC were significantly associated with 8-iso- PGF2 /Cr (Table 3). MAGE was significantly associated with 8-iso-PGF2 /Cr in all three T1DM groups. IAUC was significantly associated with 8-iso-PGF2 /Cr during the honeymoon phase (group B). HbA1c was positively associated with 8-iso-PGF2 /Cr at both the acute metabolic disturbance and long-standing phases. In the linear regression model, no collinearity existed among the independent variables.

6 doi: /jc jcem.endojournals.org 1833 Figure 2. Correlation between 8-iso-PGF2 /Cr and blood glucose fluctuation parameters. The x-axis shows MBG (A), SDBG (B), MAGE (C), and IAUC (D). The y-axis indicates the level of oxidative stress (8-iso-PGF2 /Cr). The blue, green, and red scatter plots represent acute metabolic disturbance, group A; the honeymoon period, group B; and the long-standing phase, group C, respectively. Blue, green, and red lines are bestfitting lines for each group. Discussion This study showed that the 8-iso-PGF2 /Cr, an indicator of oxidative stress, was significantly higher in children with T1DM than in children with normal glucose metabolism. This is consistent with previous studies in T2DM and T1DM patients (8, 9). The results not only support the hypothesis that hyperglycemia can increase oxidative stress (22, 23) but also indicate for the first time a positive association between glucose fluctuations and 8-iso- Table 3. PGF2 /Cr in T1DM patients. Positive associations across the different phases of T1DM add to the evidence of a relationship between glucose fluctuations and oxidative stress. In this study, the 8-iso-PGF2 /Cr progressively increased across the phases of T1DM. It was significantly higher at the long-standing phase at least 5 years after diagnosis than in newly diagnosed children and recently diagnosed honeymoon-phase children. The increase of oxidative stress level may partially be explained by the in- Association of 8-Iso-PGF2 /Cr With Indicators of Blood Glucose Levels in T1DM Patients Group A (Acute Metabolic Disturbance Phase) Group B (Honeymoon Phase) Group C (Long Standing Phase) All Groups Multivariate Analysis Standardized Coefficients P Standardized Coefficients P Standardized Coefficients P Standardized Coefficients Disease duration Age HbA1c MAGE IAUC r P

7 1834 Meng et al Glucose Fluctuations and Oxidative Stress in T1DM J Clin Endocrinol Metab, May 2015, 100(5): creased age and higher BMI among children with longstanding T1DM. However, the significance of betweengroup differences in 8-iso-PGF2 excretion remained after adjusting for gender, age, BMI, blood pressure, and lipid profiles. Additionally, a few T1DM patients are likely to be pubertal. Although puberty might be an important factor that influences oxidative stress levels, in all three T1DM groups, we did not observe significant correlations between pubertal status with 8-iso-PGF2 /Cr (r 0.024; P.41). CGM revealed that the HbA1c level was highest in children with T1DM during the acute metabolic disturbance phase. CGM also showed that MAGE and SDBG were significantly higher in children with T1DM at the acute metabolic disturbance phase than in the control children. MAGE and SDBG were lower in these newly diagnosed children than in children with a long-standing diagnosis, suggesting that glycemic excursion in the acute metabolic disturbance phase consisted of persistent hyperglycemia with moderate fluctuations. Control of HbA1c was better in children in the T1DM honeymoon phase as shown by their stable baseline blood glucose and acute postprandial increase of blood glucose. Patients in group B all received conventional two injections daily, while patients in group C took more multiple daily injections/continuous subcutaneous insulin infusions. The 8-iso-PGF2 /Cr is higher in group C than in group B. It means that more injections didn t affect the level of 8-iso-PGF2 /Cr. The long-standing phase was characterized by a high blood glucose baseline level and substantial amplitude of glucose excursions. MAGE and SDBG were the highest in this phase vs the other T1DM phases. The 8-iso-PGF2 /Cr did not change in parallel with HbA1c levels in any T1DM phase; 8-iso- PGF2 excretion may reflect the combined effect of baseline high blood glucose levels and glucose fluctuations. In the control group, glucose fluctuations were within the normal range and were not correlated with 8-iso- PGF2 /Cr. Therefore, the effects of glucose fluctuation on oxidative stress response should be considered physiological and normal. Siegelaar et al (24) did not find any correlation between glucose fluctuations and oxidative stress in well-controlled T2DM patients with a mean HbA1c of 6.9%. They proposed that only high levels of HbA1c could activate the mechanisms by which glucose fluctuations increase oxidative stress (24). Our study supports the speculation that high blood glucose levels are necessary before fluctuations can activate oxidative stress. Multiple regression analysis showed that oxidative stress levels were associated with MAGE in all three T1DM phases, suggesting that it might be a useful index to assess glucose fluctuations in children. At the honeymoon phase, IAUC was more closely associated with oxidative stress than was MAGE (0.676 vs 0.457), suggesting that postprandial hyperglycemia was a major factor associated with increased oxidative stress in those children. Ultimately, SDBG, MAGE, and IAUC may reflect differences in the nature of oxidative stress at different phases of T1DM in children. For example, clinical benefits might be achieved during the honeymoon phase by dividing meals into multiple, smaller portions given over an extended period of time or instead of a regimen of basal-bolus. Multiple linear regression analysis indicated that both chronic persistent hyperglycemia and acute glycemic excursion can activate oxidative stress during the acute metabolic disturbance and long-standing phases of T1DM. A stronger association was observed for MAGE than for HbA1c in children in these two T1DM phases; thus, acute glycemic excursion may be more effective in triggering the activation of oxidative stress than chronic persistent hyperglycemia. This finding has important clinical implications. First, rapid, substantial glucose fluctuations should be avoided, even if HbA1c is well controlled and normal; and second, glucose fluctuations are more harmful in the presence of hypoglycemia. The Diabetes Control and Complications Trial (DCCT) found a significant positive correlation between the risk of microvascular complications and glucose levels in T1DM patients, even if HbA1c levels had not changed (25, 26). The glucose fluctuation data obtained in this study may help to explain the DCCT findings. Using the glucose clamp technique, which simulates either glucose fluctuations or constant blood hyperglycemia, Ceriello et al (27) found that glucose fluctuations had a more adverse impact on endothelial function and oxidative stress. Overall, the evidence suggests that clinicians should find ways to reduce glucose fluctuations in T1DM patients, including those with well-controlled HbA1c levels. It is well known that increased activity of the polyol and hexosamine pathways, activation of protein kinase C, and increase of glycation end-products promote the development of diabetic vascular complications. Brownlee proposed that increased oxidative stress is an unifying mechanism in this process (3). Excess reactive oxygen species can damage cellular DNA and trigger the activation of polyadenylation diphosphate ribose polymerase, which in turn activates pathways leading to endothelial damage (5). Perhaps glucose fluctuations can lead to dysfunction of the mitochondrial electron transport chain that leads to increased reactive oxygen species production and induction of intracellular oxidative stress. It remains unclear why glucose fluctuations are more likely to cause oxidative damage and apoptosis than chronic hyperglycemia. Some studies suggest that glucose fluctuations can down-regulate expression of genes that regulate the decomposition of

8 doi: /jc jcem.endojournals.org 1835 free radicals (28, 29). Another explanation is that glucose fluctuations may affect the metabolic memory of vascular endothelial cells. The functioning of antioxidant pathways may allow endothelial cells to adapt to exposure to constant hyperglycemia and maintain a relatively stable extent of morphological and functional damage. In this model, exposure to an intermittent high-glucose challenge would impair the adaptation and accelerate oxidative damage (30). In addition to glucose fluctuations, we observed that duration of T1DM may have stronger association with 8-iso-PGF2 /Cr than age. This suggests that children with T1DM have increased oxidative stress levels as the duration is extending, which may not be substantially altered by age. Our study has several strengths. We enrolled only nonobese children without dyslipidemia, hypertension, coronary heart disease, and acute complications of diabetes, therefore eliminating the potentially confounding effects of those factors on oxidative stress and reducing the possibility of experimental bias in the study on the association of glucose fluctuation and oxidative stress. Second, we included staging of the natural history of T1DM in the study design. We found that there were significant associations between glucose fluctuations and oxidative stress in all three phases. The main limitation of this study is the relatively small sample size. Additional studies with larger sample sizes are needed to confirm our conclusions. In conclusion, glucose fluctuations were positively associated with oxidative stress in children with T1DM. Glucose fluctuations may have a stronger effect than sustained chronic hyperglycemia on triggering oxidative stress, but coexisting high levels of blood glucose may be required. T1DM treatment should emphasize reducing glucose fluctuations as well as achieving the overall glucose control target. Further studies are warranted to confirm our findings and to demonstrate the underlying mechanisms. Acknowledgments We thank all the doctors who worked in the endocrinology ward during this research. They were a great help. Address all correspondence and requests for reprints to: Chunxiu Gong, South Lishi Road, 56, Xicheng District, Beijing , P. R. China. chunxiugong@163.com. This work was supported by the Capital Development Fund (Grant ) of China. Author Contributions: study concept and design, C.G., B.C., X.M.; acquisition of cases, B.C., D.W., Y.G., L.W., X.L., M.L., W.L., C.S.; collection of urine samples, X.M., L.W.; data acquisition and management, X.M.; statistical analysis, X.M.; critique and review, X.P.; draft of the manuscript, X.M., C.G.; critical revision of the manuscript for important intellectual content, C.G.; and study supervision, C.G. Disclosure Summary: The authors have nothing to disclose. References 1. Brownlee M, Hirsch IB. Glycemic variability: a hemoglobin A1cindependent risk factor for diabetic complications. JAMA. 2006; 295: Šoupal J, Škrha J Jr, Fajmon M, et al. Glycemic variability is higher in type 1 diabetes patients with microvascular complications irrespective of glycemic control. Diabetes Technol Ther. 2014;16: Brownlee M. The pathobiology of diabetic complications: a unifying mechanism. Diabetes. 2005;54: Zoppini G, Verlato G, Targher G, et al. Is fasting glucose variability a risk factor for retinopathy in people with type 2 diabetes? Nutr Metab Cardiovasc Dis. 2009;19: 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: El-Osta A, Brasacchio D, Yao D, et al. Transient high glucose causes persistent epigenetic changes and altered gene expression during subsequent normoglycemia. J Exp Med. 2008;205: Quagliaro L, Piconi L, Assaloni R, et al. Intermittent high glucose enhances ICAM-1, VCAM-1 and E-selectin expression in human umbilical vein endothelial cells in culture: the distinct role of protein kinase C and mitochondrial superoxide production. Atherosclerosis. 2005;183: Monnier L, Mas E, Ginet C, et al. Activation of oxidative stress by acute glucose fluctuations compared with sustained chronic hyperglycemia in patients with type 2 diabetes. JAMA. 2006;295: Wentholt IM, Kulik W, Michels RP, Hoekstra JB, DeVries JH. Glucose fluctuations and activation of oxidative stress in patients with type 1 diabetes. Diabetologia. 2008;51: Monnier L, Colette C, Mas E, et al. Regulation of oxidative stress by glycaemic control: evidence for an independent inhibitory effect of insulin therapy. Diabetologia. 2010;53: Monnier L, Colette C, Michel F, Cristol JP, Owens DR. Insulin therapy has a complex relationship with measure of oxidative stress in type 2 diabetes: a case for further study. Diabetes Metab Res Rev. 2011;27: Frontoni S, Di Bartolo P, Avogaro A, Bosi E, Paolisso G, Ceriello A. Glucose variability: an emerging target for the treatment of diabetes mellitus. Diabetes Res Clin Pract. 2013;102: Manohar C, Levine JA, Nandy DK, et al. The effect of walking on postprandial glycemic excursion in patients with type 1 diabetes and healthy people. Diabetes Care. 2012;35: Morrow JD, Frei B, Longmire AW, et al. Increase in circulating products of lipid peroxidation (F2-isoprostanes) in smokers. Smoking as a cause of oxidative damage. N Engl J Med. 1995;332: Stojiljkovic MP, Lopes HF, Zhang D, Morrow JD, Goodfriend TL, Egan BM. Increasing plasma fatty acids elevates F2-isoprostanes in humans: implications for the cardiovascular risk factor cluster. J Hypertens. 2002;20: Davi G, Alessandrini P, Mezzetti A, et al. In vivo formation of 8-Epiprostaglandin F2 is increased in hypercholesterolemia. Arterioscler Thromb Vasc Biol. 1997;17: Keaney JF Jr, Larson MG, Vasan RS, et al. Obesity and systemic oxidative stress: clinical correlates of oxidative stress in the Framingham Study. Arterioscler Thromb Vasc Biol. 2003;23:

9 1836 Meng et al Glucose Fluctuations and Oxidative Stress in T1DM J Clin Endocrinol Metab, May 2015, 100(5): Craig ME, Jefferies C, Dabelea D, Balde N, Seth A, Donaghue KC. Definition, epidemiology, and classification of diabetes in children and adolescents. Pediatr Diabetes. 2014;15 Suppl 20: Mortensen HB, Hougaard P, Swift P, et al. New definition for the partial remission period in children and adolescents with type 1 diabetes. Diabetes Care. 2009;32: Mezzetti A, Cipollone F, Cuccurullo F. Oxidative stress and cardiovascular complications in diabetes: isoprostanes as new markers on an old paradigm. Cardiovasc Res. 2000;47: Morrow JD. Quantification of isoprostanes as indices of oxidant stress and the risk of atherosclerosis in humans. Arterioscler Thromb Vasc Biol. 2005;25: Ceriello A, Kumar S, Piconi L, Esposito K, Giugliano D. Simultaneous control of hyperglycemia and oxidative stress normalizes endothelial function in type 1 diabetes. Diabetes Care. 2007;30: Esposito K, Nappo F, Marfella R, et al. Inflammatory cytokine concentrations are acutely increased by hyperglycemia in humans: role of oxidative stress. Circulation. 2002;106: Siegelaar SE, Barwari T, Kulik W, Hoekstra JB, DeVries JH. No relevant relationship between glucose variability and oxidative stress in well-regulated type 2 diabetes patients. J Diabetes Sci Technol. 2011;5: The relationship of glycemic exposure (HbA1c) to the risk of development and progression of retinopathy in the Diabetes Control and Complications Trial. Diabetes. 1995;44: White NH, Sun W, Cleary PA, et al. Prolonged effect of intensive therapy on the risk of retinopathy complications in patients with type 1 diabetes mellitus: 10 years after the Diabetes Control and Complications Trial. Arch Ophthalmol. 2008;126: 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: Meugnier E, Faraj M, Rome S, et al. Acute hyperglycemia induces a global downregulation of gene expression in adipose tissue and skeletal muscle of healthy subjects. Diabetes. 2007;56: Schisano B, Tripathi G, McGee K, McTernan PG, Ceriello A. Glucose oscillations, more than constant high glucose, induce p53 activation and a metabolic memory in human endothelial cells. Diabetologia. 2011;54: Azuma K, Kawamori R, Toyofuku Y, et al. Repetitive fluctuations in blood glucose enhance monocyte adhesion to the endothelium of rat thoracic aorta. Arterioscler Thromb Vasc Biol. 2006;26: