The Role of Markers of Low-Grade Inflammation for the Early Time Course of Glycemic Control, Glucose Disappearance Rate, and β-Cell Function in Recently Diagnosed Type 1 and Type 2 Diabetes

Hyperglycemia is the primary diagnostic criterion of diabetes. The most abundant types of diabetes, type 1 and type 2, differ in etiology (1) but share the common feature of impaired β-cell function before the onset of the disease (2,3). β-Cell function generally declines rapidly in type 1 diabetes due to progressive autoimmune destruction (4), whereas it deteriorates more slowly before type 2 diabetes (3). Genome-wide association studies have emphasized the role of impaired β-cell function for type 2 diabetes (5). Nevertheless, chronic insulin resistance is considered more important in triggering type 2 diabetes (6). Of note, data also suggest an accelerating role of insulin resistance for type 1 diabetes (7).

Inflammatory processes can contribute to both insulin resistance and β-cell dysfunction. In type 2 diabetes, most studies have focused on low-grade inflammation as a trigger for insulin resistance. However, data also link inflammation and β-cell function in type 2 diabetes (8). In type 1 diabetes, inflammatory processes appear more relevant because of their β-cell toxicity rather than their impact on insulin resistance (9). Given the importance of insulin resistance and β-cell function for near-normoglycemic control, whether circulating inflammation-related biomarkers indeed associate with insulin resistance and β-cell function during the progression of type 1 and type 2 diabetes is relevant. Associations between biomarkers of low-grade inflammation and diabetes status, insulin resistance, and β-cell function have been analyzed in cross-sectional studies (10,11). In addition, prospective studies have examined the role of acute phase proteins, cytokines, chemokines, and biomarkers of endothelial function in the context of incident diabetes, mostly type 2 diabetes (12–14). In contrast, data on the impact of inflammation-related biomarkers on changes of insulin resistance and β-cell function in the early course of the disease are scarce (15).

It is conceivable that inflammatory processes affect insulin sensitivity and β-cell function not only before but also after the clinical diagnosis of type 1 and type 2 diabetes. Thus, this study aimed to investigate the associations of 2-year changes of markers of low-grade inflammation with 2-year changes of glycemic control, β-cell function, and glucose disappearance rate in patients with recently diagnosed type 1 or type 2 diabetes. We hypothesized that higher levels of biomarkers of low-grade inflammation are associated with the early progression of recently diagnosed diabetes and aimed to describe the impact of changes in the levels of biomarkers of low-grade inflammation on early changes of glycemic control, β-cell function, and glucose disappearance rate in both type 1 and type 2 diabetes.

The German Diabetes Study (GDS) is a prospective observational study investigating the natural history of recent-onset diabetes and the development of diabetes-associated complications to improve risk assessment and targeted treatment of patients with diabetes (16,17). The primary outcome measure of the GDS is the change in insulin sensitivity; secondary outcome measures are change of insulin secretion and incidence of diabetes-related complications. The study is performed according to the Declaration of Helsinki and is approved by the ethics committee of Heinrich Heine University Düsseldorf. All patients give written informed consent before inclusion into the study. The study was started in September 2005 and still includes new patients, aiming for 2,000 in total. For this substudy, all patients were included for whom baseline and 2-year follow-up data were available. For this group, 200 baseline examinations were conducted between September 2005 and December 2008 and follow-up examinations between November 2007 and October 2011. Baseline inclusion criteria were age between 18 and 69 years and known diabetes duration <12 months. Exclusion criteria were diabetes owing to secondary diseases, pregnancy, acute infections, immunosuppressive therapy, evidence of congestive heart failure, kidney diseases, liver diseases, symptomatic peripheral arterial disease, psychiatric or addictive diseases, history of malignancies, or participation in pharmacological intervention studies. Patients with previous gestational diabetes mellitus with type 2 diabetes after postpartum normalization of glucose tolerance are not excluded from this study. Patients are referred to the German Diabetes Center by their local general practitioner or by specialized outpatient units or recruited by advertisement. If necessary, diabetes diagnosis is confirmed by records of plasma glucose ≥7.0 mmol/L at fasting or ≥11.1 mmol/L at 120 min of a standardized 75-g oral glucose tolerance test in accordance with American Diabetes Association guidelines (1). Type 1 diabetes is defined by the presence of autoantibodies, classic symptoms, and/or ketoacidosis at onset of disease.

Patients underwent detailed metabolic phenotyping at baseline and 2-year follow-up examination after fasting overnight (≥8 h) and after having stopped oral glucose-lowering medication for 3 days. Patients on insulin treatment administered their last insulin dose in the evening before the first examination day. On the first day, patients underwent anthropometric examinations, fasting blood analyses, and glucagon stimulation testing. A structured interview covering medical history, family history of chronic diseases, and socioeconomic status was conducted. This was followed by an intravenous glucose tolerance test (IVGTT) on the next day.

BMI was calculated from body weight and height recorded using a calibrated scale with stadiometer (seca 674; Hamburg, Germany) without wearing shoes and in light underwear to the nearest 0.5 kg and 0.1 cm, respectively.

A catheter was inserted into an antecubital vein, and a fasting blood sample was drawn. At 0 min, a bolus of 1 mg glucagon (GlucaGen; Novo Nordisk, Mainz, Germany) was injected intravenously within 60 s. At 6 min, a second blood sample was obtained (18). The ratio between C-peptide concentrations at 6 min and 0 min was calculated to assess glucagon-stimulated C-peptide secretion capacity.

The protocol of the IVGTT has been previously described (19). The acute C-peptide glucose-dependent response (ACPRG) was calculated as the average of C-peptide concentrations between 2 and 10 min following glucose injection. This value is an extension to the C-peptide of the acute insulin response introduced by Kahn et al. (20). It represents the β-cell direct secretory response to glucose stimulation. Glucose-stimulated total incremental area under the curve (AUC) for C-peptide (iAUC_CP) was calculated using the trapezoidal rule subtracting the basal area from the AUC. β-Cell function was also characterized by calculating the incremental prehepatic β-cell function (iBCF) based on modeling analysis of C-peptide to assess prehepatic insulin secretion (21). Glucose disappearance rate is a measure of intravenous glucose tolerance and represents the net glucose elimination rate after glucose injection. It was calculated as the slope for the interval 6–40 min after glucose injection using log-transformed glucose values (22).

Supplementary Table 1 provides detailed information with respect to biospecimen type, storage, freeze-thaw cycles, intra- and interassay coefficients of variation, and measurement ranges for all analytes (HbA_1c, glucose, C-peptide, lipids, high-sensitivity C-reactive protein [hsCRP], IL-6, IL-18, soluble E-selectin [sE-selectin], soluble intercellular adhesion molecule-1 [sICAM-1], and autoantibodies). For all parameters, blood samples at baseline and follow-up were taken using the same tubes and tube additives. Immune mediator measurements of samples from baseline and follow-up were done for all samples at the same time, using the same instruments and methods.

Results are presented as mean ± SD for normally distributed data and median (25th; 75th percentiles) for nonnormally distributed data. Differences between baseline and 2-year follow-up were assessed using paired Student t test for normally distributed continuous variables, Wilcoxon signed rank test for nonnormally distributed continuous variables, McNemar test for paired samples for categorical variables with two characteristics, and Bowker test of symmetry for categorical variables with more than two characteristics.

Multivariable linear regression models were used to analyze the associations between 2-year changes (2-year follow-up divided by baseline) of markers of inflammation (independent variable) with 2-year changes of glycemic control, β-cell function, and glucose disappearance rate (dependent variable). Glucagon-stimulated C-peptide secretion capacity, ACPRG, iAUC_CP, iBCF, hsCRP, IL-6, IL-18, sE-selectin, and sICAM-1 were not normally distributed and, thus, were entered into the models as ln-transformed variables. In the basic model (model 1), the value of the respective measure of glucose metabolism at baseline, the concentration of the marker of low-grade inflammation at baseline, age at baseline, sex, and diabetes type were considered as covariables potentially affecting the associations. A second model (model 2) was additionally adjusted for BMI at baseline, change of BMI, change of the type of glucose-lowering medication, smoking status at baseline, and change of smoking status. The fit of the full models was checked for normality using Q-Q plots. Residuals were checked for homoscedasticity. Scatter plots of the predicted value with the dependent variable were evaluated. Additionally, R² values and P values of the F statistic were calculated for each model (Supplementary Tables 2 and 3). Regression coefficients (β) should be interpreted as follows: A doubling of the ratio of hsCRP, IL-6, IL-18, sE-selectin, and sICAM-1 between baseline and 2-year follow-up associates with an absolute change of HbA_1c and glucose disappearance rate within the first 2 years by β, and a doubling of the ratio of 2-year follow-up/baseline values of hsCRP, IL-6, IL-18, sE-selectin, and sICAM-1 associates with a percent change of C-peptide secretion capacity, ACPRG, iAUC_CP, and iBCF within the first 2 years by β. Post hoc power calculations and patient numbers needed to detect changes were conducted as described in the Supplementary Data (Supplementary Tables 4–7).

Additionally, four sensitivity analyses were conducted. First, glucose disappearance rate was added as a potential confounder to model 2 for those models with indices of insulin secretion as a dependent variable. Second, β-values of patients with type 1 and type 2 diabetes were tested for differences as described (23). Third, only patients without glucose-lowering medications at baseline and follow-up were included in linear regression models (Supplementary Tables 8 and 9). Fourth, disease status of bronchial asthma and chronic obstructive pulmonary disease (COPD) at baseline (presence of the disease within the past 12 months) and follow-up (presence of the disease within the past 24 months) were added as potential confounders to model 2. P < 0.05 were considered statistically significant. All statistical analyses were carried out using SAS version 9.2 software (SAS Institute, Cary, NC).

A total of 137 individuals with type 1 and type 2 diabetes participated in the baseline and 2-year follow-up examinations. Follow-up time was 2.1 ± 0.1 years. The response rate for the follow-up was 68.5% (Fig. 1 and Supplementary Fig. 1). Patients with type 2 diabetes who were lost to follow-up were younger and had shorter known diabetes duration compared with those who participated in both examinations. Patients with type 1 diabetes who quit study participation did not differ from those who continued (Supplementary Table 10). Type 1 diabetes was present in 31% of the study population (Table 1). At baseline, all patients with type 2 diabetes were autoantibody negative. Of the patients with type 1 diabetes, 15% were autoantibody negative, 36% had one autoantibody, and 49% expressed two autoantibodies. Patients exhibited good glucometabolic control. Between baseline and 2-year follow-up, glucagon-stimulated C-peptide secretion and sE-selectin increased, whereas ACPRG and iAUC_CP decreased in patients with type 2 diabetes. In type 1 diabetes, BMI and sICAM-1 increased, whereas ACPRG, iAUC_CP, and glucose disappearance rate decreased within the first 2 years after diabetes diagnosis (Table 1).

In patients with type 2 diabetes, increases of serum concentrations of hsCRP, sE-selectin, and sICAM-1 (per doubling of the ratio of 2-year follow-up/baseline values) within the first 2 years after diagnosis was associated with absolute increases of HbA_1c by 0.46%, 0.90%, and 1.69%, respectively (model 2, all P < 0.01). No association was observed between changes of IL-6 or IL-18 and changes of HbA_1c. Additionally, 2-year changes of sE-selectin were associated with a relative decrease of iBCF of 34.87% (model 2). The associations between 2-year changes of sE-selectin and sICAM-1 with HbA_1c and of sE-selectin with iBCF were not explained by serum concentrations of triglycerides, HDL cholesterol, or LDL cholesterol at baseline (data not shown). Associations between changes of hsCRP and sE-selectin with changes of C-peptide secretion capacity from glucagon stimulation testing lost significance after adjustment (model 2). Associations between changes of IL-6 and changes of iAUC_CP as well as between sICAM-1 and ACPRG were only significant in the fully adjusted model (model 2). There was no evidence for associations between changes of serum concentrations of biomarkers of low-grade inflammation and changes of glucose disappearance rate (Table 2 and Supplementary Table 2).

In patients with type 1 diabetes, increases of serum concentrations of IL-18 (per doubling of the ratio of 2-year follow-up/baseline values) within the first 2 years after diagnosis associated with absolute increases of HbA_1c by 1.18% after multiple adjustment (model 2). Associations between changes of sE-selectin with changes of C-peptide secretion capacity from glucagon stimulation testing lost significance after adjustment for confounders (model 2). There was no evidence that changes of biomarkers of subclinical inflammation associated with changes of ACPRG, iAUC_CP, iBCF, or glucose disappearance rate in patients with type 1 diabetes (Table 3 and Supplementary Table 3).

Results virtually did not change when we also adjusted the data for glucose disappearance rate (data not shown). COPD and bronchial asthma were present in ≤5% of patients with the only exception being of asthma in 10% and 12% of patients with type 1 diabetes at baseline and follow-up, respectively (Table 1). Because both diseases are important determinants of inflammatory status, we included disease status at baseline and follow-up as potential confounding factors to linear regression models. Again, results were virtually unchanged (data not shown).

Because associations between changes of markers of low-grade inflammation with glycemic control, β-cell function, and glucose disappearance rate might already differ between type 1 and type 2 diabetes during early disease progression, we also tested β-values for differences between diabetes types. The association between 2-year changes of sE-selectin and C-peptide secretion capacity from glucagon stimulation testing was different between diabetes types (type 1 diabetes β = 46.10, type 2 diabetes β = −13.29, P = 0.007 [model 1]). There was no indication for any other statistically significant difference between diabetes types.

We also repeated regression analyses with only patients with type 2 diabetes who had no glucose-lowering medication at baseline and follow-up. Of the associations that were significant in the main analysis, the association between changes of hsCRP and sICAM-1 with changes of HbA_1c remained significant. In the subgroup of patients without glucose-lowering medication, additional associations were found between changes of IL-6 with changes of HbA_1c and changes of hsCRP with changes of iBCF. Associations were attenuated and lost significance in this subgroup between sE-selectin and HbA_1c as well as iBCF (Supplementary Tables 8 and 9).

This study in recently diagnosed patients with type 1 and type 2 diabetes with overall good metabolic control demonstrates that 2-year increases of hsCRP, sE-selectin, and sICAM-1 in type 2 diabetes and of IL-18 in type 1 diabetes associated with 2-year deterioration in glycemic control (i.e., increases in HbA_1c). Furthermore, 2-year increases of sE-selectin were related to 2-year impairment of glucose-stimulated insulin secretion (i.e., decreases of iBCF). These associations were not explained by baseline BMI or concomitant BMI changes.

Changes of hsCRP, sE-selectin, and sICAM-1 within the first 2 years after diabetes diagnosis positively associated with 2-year changes of HbA_1c in type 2 diabetes. Changes of IL-18 were positively related to changes of HbA_1c in type 1 diabetes, extending the current literature of cross-sectional studies (24–28) because the present data provide prospective associations between glycemic control and markers of low-grade inflammation. Additionally, unlike previous investigations in patients with longer diabetes duration (29–31), we included patients with recently diagnosed diabetes.

The present data are in line with cross-sectional studies reporting associations of higher serum concentrations of hsCRP, sE-selectin, and sICAM-1 with higher HbA_1c (26,27). The associations were confirmed in prospective studies (15,30,31). However, two of these prospective studies did not focus on recently diagnosed diabetes (30,31). Concerning hsCRP and HbA_1c, the one study considered associations between hsCRP at one time point with the average HbA_1c during follow-up (31), and another compared average hsCRP during the study period with changes of HbA_1c (15). High IL-18 levels were also previously shown to be related to acute hyperglycemia as well as to long-term glycemic control in cross-sectional analyses of healthy people, individuals with impaired glucose tolerance, and patients with type 1 diabetes with a mean duration of 9 years (24,28). Correction of hyperglycemia to normoglycemia in patients with recently diagnosed type 2 diabetes resulted in normalization of increased plasma concentrations of sICAM-1 (25). Furthermore, reductions in hsCRP levels were associated with reductions in HbA_1c in patients with type 2 diabetes on structured self-monitoring of blood glucose (29). Both associations (changes of hsCRP and sICAM-1 with changes of HbA_1c) could be replicated in the present patients with type 2 diabetes not taking glucose-lowering medications. This indicates that oral glucose-lowering medication or insulin does not appear to affect these relations, whereas associations of sE-selectin and IL-6 with HbA_1c seem to be influenced more strongly by glucose-lowering medication.

Associations between markers of low-grade inflammation and HbA_1c might be mediated by glucose-induced oxidative stress (26,28) and thereby could contribute to early stages of atherogenesis (25). Additionally, HbA_1c has been related to reduced antioxidant trapping capacity in patients with type 2 diabetes (30). However, whether these early changes increase the risk for chronic vascular complications in diabetes remains to be proven (25).

In the present cohort, changes of sE-selectin inversely associated with changes of iBCF in patients with type 2 diabetes. In the subgroup of patients with type 2 diabetes without glucose-lowering medication, this relation was attenuated probably owing to power issues. Additionally, changes of IL-6 associated with changes of iBCF in these patients. The few available data on the association between soluble adhesion molecules and β-cell function are from cross-sectional studies. In healthy men, lipid infusion increases concentrations of circulating adhesion molecules (32). In other studies, sICAM-1 concentrations were also associated with the AUC_insulin and with maximal insulin secretion after the ingestion of a lipid-rich meal (33). Circulating concentrations of sE-selectin, but not sICAM-1, were related to AUC_insulin in patients with impaired glucose tolerance (34) as well as with fasting insulin concentrations in patients with type 2 diabetes (27). By proposing an association between soluble adhesion molecules, especially sE-selectin, and the early development of insulin secretion and β-cell function in patients with diabetes, the present results provide new insights into a possible link among vascular inflammation, endothelial activation, and decreasing β-cell function. Increased concentrations of soluble adhesion molecules reflecting endothelial activation were previously described in the state of hyperinsulinemia and insulin resistance (34). Because elevated circulating lipid concentrations impair endothelial function (35), we also tested whether serum concentrations of triglycerides, HDL cholesterol, and LDL cholesterol influenced the findings. However, the results indicate that the associations between soluble adhesion molecules with glycemic control and β-cell function were not confounded by serum lipid levels.

In the present cohort, there was no evidence that changes of biomarkers of low-grade inflammation were associated with changes of glucose disappearance rate, a measure of intravenous glucose tolerance. Therefore, the data differ from previous findings from cross-sectional studies that reported associations between elevated concentrations of hsCRP and insulin resistance in patients with recent-onset type 1 diabetes (15) as well as associations between sE-selectin and insulin resistance in patients with type 2 diabetes (27,36). Relations between sE-selectin and insulin resistance were also observed in healthy individuals and those with impaired glucose tolerance (10,34,37). In contrast, the present data suggest that the biomarkers of low-grade inflammation analyzed in this study do not play a key role in the deterioration of the glucose disappearance rate during the early time course of type 1 and type 2 diabetes. Given positive findings from prospective studies on the associations between these biomarkers and incident diabetes (12–14), we believe that this question merits further studies in larger cohorts.

Although mechanisms leading to type 1 and type 2 diabetes partially differ (1), the impact of low-grade inflammation on the deterioration of glycemia may be comparable between both diabetes types, as suggested by the present results. Regarding all findings on associations between biomarkers of low-grade inflammation with HbA_1c, insulin secretion, and β-cell function, sensitivity analyses only revealed diabetes-specific patterns for the relation between changes of hsCRP with changes of C-peptide secretion from glucagon stimulation testing. Evidence suggests that biomarkers of inflammation are similar risk factors for diabetes complications (e.g., cardiovascular diseases) in both type 1 and type 2 diabetes (30,31).

Strengths of this study are the use of prospective data on biomarkers of low-grade inflammation and on the natural history of diabetes. Additionally, the study comprises patients with type 1 and type 2 diabetes with comparable short-term duration of disease and similar glycemic control who also underwent extensive metabolic characterization. We followed the same protocols and standard operating procedures during the baseline and follow-up examinations with respect to blood sampling and biomarker measurement to minimize the impact of any preanalytical, interassay, or biological/environmental variations on the results. Furthermore, the study population is unique regarding patients with type 1 diabetes because we only included those in whom diabetes developed at an age of at least 17 years. On the other hand, this method restricts comparability to childhood and adolescence disease onset and limits the generalizability of the results. This can be exemplified by the 15% of patients with type 1 diabetes who are autoantibody negative at baseline. A rather mild disease progression might be characteristic at least for some of the patients with adult-onset type 1 diabetes in this cohort.

There are also several limitations. First, lack of genetic screening exists for maturity-onset diabetes of the young or for even less frequent rare forms of diabetes. However, according to the SEARCH for Diabetes in Youth cohort, the prevalence for a mutation in one of the three maturity-onset diabetes of the young genes is only 1.2% (38). Second, we obtained the glucose disappearance rate as a measure of intravenous glucose tolerance from the IVGTT, although the hyperinsulinemic-euglycemic clamp is considered the gold standard method for assessment of insulin sensitivity. Third, HbA_1c was measured using the Diabetes Control and Complications Trial (DCCT) method instead of the newer World Health Organization–recommended International Federation of Clinical Chemistry and Laboratory Medicine method. Fourth, the good glucometabolic control of the patients might have originated from selection bias (i.e., the higher interest of health-conscious people in clinical study participation); thus, transferability of the findings to the general population of patients with type 1 and type 2 diabetes requires further testing. Fifth, regression models, especially the fully adjusted model (model 2), for type 1 diabetes might be overfitted because of the small sample size. Additionally, when considering multiple testing and applying Bonferroni correction to results from patients with type 1 and type 2 diabetes, only the associations between sICAM-1 and HbA_1c in type 2 diabetes remains significant in both models (P < 0.05/60). Based on the R² values, the strongest relationships were observed for IVGTT-derived C-peptide secretion and β-cell function (ACPRG, iAUC_CP, iBCF) in patients with type 1 and type 2 diabetes. Weak associations were found for glucose disappearance rate in type 2 diabetes and for HbA_1c in type 1 diabetes (Supplementary Tables 2 and 3). We cannot exclude that the weak association for HbA_1c might be a result of the variability of the DCCT method. Along with that, the present analyses might be underpowered, again specifically applying to type 1 diabetes, which implies the possibility of type II error. However, for some of the analyses, either associations in the cohort or associations in general appear to be absent when assuming a clinically relevant change of the dependent variable with a doubling of the low-grade inflammation biomarker (both indicated by a patient number >10,000 in Supplementary Tables 4–7). Additionally, no real association instead of type II error is possible if power calculations yield a large number of patients needed to detect significant differences (Supplementary Tables 4 and 5).

Low-grade inflammation may negatively affect glycemic control, first-phase insulin secretion, and β-cell function in the early time course of type 1 and type 2 diabetes. Long-term follow-up of patients with repetitive measurements over time will allow us to examine the extent to which the associations observed in early disease development will influence the progression of diabetes and the occurrence of macro- and microvascular diabetes-related complications at later stages of the disease.

GDS Group. M. Roden (speaker), A.E. Buyken, J. Eckel, G. Geerling, H. Al-Hasani, C. Herder, A. Icks, J. Kotzka, O. Kuss, E. Lammert, J. Lundbom, K. Müssig, P. Nowotny, W. Rathmann, J. Szendroedi, D. Ziegler, and their coworkers who are responsible for the design and performance of the GDS.

Acknowledgments. The authors appreciate the voluntary contribution of all study participants. They also thank the staff of the Clinical Research Center of the German Diabetes Center (DDZ) for excellent technical assistance and taking care of the patients. The authors further thank Ulrike Partke and Gabi Gornitzka, DDZ, for technical assistance.

Funding. The GDS was initiated and is performed by the DDZ, which is funded by the German Federal Ministry of Health and the Ministry of Innovation, Science, Research and Technology of the State North Rhine-Westphalia. This study was supported in part by a grant from the German Federal Ministry of Education and Research (BMBF) to the German Center for Diabetes Research (DZD e.V.) and by a grant of the Helmholtz Alliance Imaging and Curing Environmental Metabolic Diseases (ICEMED).

Duality of Interest. No potential conflicts of interest relevant to this article were reported.

Author Contributions. K.S.W. researched data and contributed to the statistical analysis and writing and critical review of the manuscript. B.N. researched data and critically reviewed the manuscript. K.S. researched data and contributed to the statistical analysis and critical review of the manuscript. G.P. assessed metabolic parameters by running the mathematical models, interpreted the results, and contributed to the writing and critical review of the manuscript. K.M. and J.S. researched data and contributed to discussion and critical review and editing of the manuscript. C.H. researched data, analyzed the biomarkers of low-grade inflammation, and contributed to the study design, discussion, and critical review and editing of the manuscript. M.R. contributed to the study design, discussion, and critical review and editing of the manuscript. M.R. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Prior Presentation. Parts of this study were presented in abstract form at the 75th Scientific Sessions of the American Diabetes Association, Boston, MA, 5–9 June 2015; the Diabetes Congress of the German Diabetes Association, Berlin, Germany, 13–16 May 2015; and the 58th Symposium of the German Society of Endocrinology, Lübeck, Germany, 18–21 March 2015.