The INNODIA Type 1 Diabetes Natural History Study: a European cohort of newly diagnosed children, adolescents and adults.

AIMS/HYPOTHESIS: Type 1 diabetes is an heterogenous condition. Characterising factors explaining differences in an individual's clinical course and treatment response will have important clinical and research implications. Our aim was to explore type 1 diabetes heterogeneity, as assessed by clinical characteristics, autoantibodies, beta cell function and glycaemic outcomes, during the first 12 months from diagnosis, and how it relates to age at diagnosis. METHODS: Data were collected from the large INNODIA cohort of individuals (aged 1.0-45.0 years) newly diagnosed with type 1 diabetes, followed 3 monthly, to assess clinical characteristics, C-peptide, HbA1c and diabetes-associated antibodies, and their changes, during the first 12 months from diagnosis, across three age groups: <10 years; 10-17 years; and ≥18 years. RESULTS: The study population included 649 individuals (57.3% male; age 12.1±8.3 years), 96.9% of whom were positive for one or more diabetes-related antibodies. Baseline (IQR) fasting C-peptide was 242.0 (139.0-382.0) pmol/l (AUC 749.3 [466.2-1106.1] pmol/l × min), with levels increasing with age (p<0.001). Over time, C-peptide remained lower in participants aged <10 years but it declined in all age groups. In parallel, glucose levels progressively increased. Lower baseline fasting C-peptide, BMI SD score and presence of diabetic ketoacidosis at diagnosis were associated with lower stimulated C-peptide over time. HbA1c decreased during the first 3 months (p<0.001), whereas insulin requirement increased from 3 months post diagnosis (p<0.001). CONCLUSIONS/INTERPRETATION: In this large cohort with newly diagnosed type 1 diabetes, we identified age-related differences in clinical and biochemical variables. Of note, C-peptide was lower in younger children but there were no main age differences in its rate of decline.

Lipid levels in a cohort of healthy Danish schoolchildren ages 5 to 17 years.

It is internationally recognized to use clinical decision limits (CDL) when interpreting the lipid levels in both adults and children, even though the evidence for children is scarce. The purpose of this study is to describe how lipid levels progress in healthy Danish children ages 5 to 17 years. This study is based on the Childhood Health, Activity, and Motor Performance School Study Denmark (CHAMPS-study DK) consisting of 1456 observations of schoolchildren aged 5 to 17 years. Participants have been tested for blood levels of total cholesterol, low-density lipoprotein (LDL), high-density lipoprotein (HDL), triglycerides, and remnant cholesterol levels are calculated. Finally, sex-specific percentile reference curves are presented. Percentile reference curves stratified by sex were generated for all cholesterols and showed that the total cholesterol level peaks at 4.32 mmol/l in 10-year-old boys and 4.46 mmol/l in nine-year-old girls. HDL levels in boys peak at 1.72 mmol/l in nine-year-old boys. HDL levels in girls and LDL levels in both sexes are nearly constant. Triglycerides kept rising to the age of 17 years in both sexes and remnant cholesterol decreased from age 5 to 17 years in both sexes. BMI z-score adjustment revealed no significant association with total cholesterol in both sexes but a significant association between HDL, LDL, triglycerides, and remnant cholesterol. This study is the first to generate percentile reference curves for blood levels of total cholesterol, LDL, HDL, triglycerides, and remnant cholesterol in a cohort of healthy Danish children aged 5 to 17 years.

Differences in insulin sensitivity in the partial remission phase of childhood type 1 diabetes; a longitudinal cohort study.

AIMS: Studies suggest that type 1 diabetes (T1D) contributes to impaired insulin sensitivity (IS). Most children with T1D experience partial remission but the knowledge regarding the magnitude and implications of impaired IS in this phase is limited. Therefore, we investigate the impact of IS on the partial remission phase. METHODS: In a longitudinal study of children and adolescents, participants were seen at three clinical visits during the first 14.5 months after diagnosis of T1D. Partial remission was defined as IDAA1c (HbA1c (%) + 4*daily insulin dose) ≤ 9. Beta-cell function was considered significant by a stimulated c-peptide > 300 pmol/L. Participants were characterized by (i) remission or non-remission and (ii) stimulated c-peptide levels above or below 300 pmol/L. IS, body mass index (BMI), total body fat, sex, age, pubertal status and ketoacidosis at onset were compared. RESULTS: Seventy-eight children and adolescents aged 3.3-17.7 years were included. At 14.5 months post-diagnosis, 54.5% of the participants with stimulated c-peptide > 300 pmol/L were not in partial remission. Participants not in remission had significant lower IS 2.5 (p = 0.032), and 14.5 (p = 0.022) months after diagnosis compared to participants in partial remission with similar c-peptide levels. IS did not fluctuate during the remission phase. CONCLUSIONS: A number of children and adolescents have impaired IS in the remission phase of paediatric T1D and are not in remission 14.5 months after diagnosis despite stimulated c-peptide > 300 pmol/L.

The metabolic syndrome is frequent in children and adolescents with type 1 diabetes compared to healthy controls.

OBJECTIVE: There is a rise in overweight and obesity among children and adolescents with type 1 diabetes (T1D) in parallel with the rise in the metabolic syndrome (MetS) among children and adolescents. The aim of the study was to describe the prevalence and characteristics of MetS in children and adolescents with T1D compared to their healthy counterparts. RESEARCH DESIGN AND METHODS: The study includes two Danish cohorts; (i) the Copenhagen cross sectional cohort 2016 of 277 children and adolescents with T1D that attend the pediatric outpatient clinic at a large hospital in greater Copenhagen and (ii) the CHAMPS-study DK which is a population-based cohort study of Danish children and adolescents (control cohort). Participants were categorized to have MetS if at least two of the following criteria were met: (i) systolic and/or diastolic blood pressure ≥ 90th percentile, (ii) waist circumference ≥90th percentile, and (iii) triglyceride ≥90th percentile and/or HDL ≤10th percentile. RESULTS: The prevalence of children with Mets in the T1D cohort was higher than in the control cohort (p = 0.002). Moreover, participants with T1D had MetS at a lower level of BMI (p < 0.001) and waist circumference (p < 0.001) than participants with MetS from the control cohort (z-scores = 0.90 and 1.51). Participants with MetS were younger than the other T1D participants (median 12.8 [9.9,14.8] vs. median 14.6 [11.2,16.9] years, p = 0.006). CONCLUSIONS: Children and adolescents with T1D have an increased risk of MetS compared to healthy controls and clinicians and caretakers should consider early prevention and health promotion strategies.

Reference serum percentile values of adiponectin, leptin, and adiponectin/leptin ratio in healthy Danish children and adolescents.

The adipokines adiponectin and leptin play key roles in human metabolic regulation and have gained great attention as biomarkers for various metabolic pathologies. Though, pediatric reference values are few and needed. This study aims to establish age- and sex-specific adipokine reference percentiles based on healthy Danish school children. Further, it elucidates sex-specific differences in associations between z-scores of examined adipokines and metabolic variables. Serum adiponectin and serum leptin from 853 observations of healthy Danish schoolchildren aged 8-17 years (median 10.0) were quantified by immunoassays. Age- and sex-specific adipokine reference percentiles were calculated cross-sectionally using the LMS method, and adipokine z-scores were calculated from the fitted model. Multiple linear regression models were used to examine sex-specific differences in associations between adipokine z-scores and various metabolic variables. Girls had a higher median value of adiponectin (11.31 vs. 10.65 µg/mL, p < .001) and leptin (2.30 vs. 1.00 ng/mL, p < .001) and a lower median value of adiponectin/leptin ratio (4.64 vs. 10.76, p < .001) compared to boys. Sex-specific differences were found in associations between adiponectin z-score and HDL (p = .010), between leptin z-score and waist circumference z-score (p = .027) and LDL (p = .048), and between adiponectin/leptin ratio z-scores and waist circumference z-score (p = .044) and LDL (p = .040). Reference percentiles of adiponectin, leptin, and adiponectin/leptin ratio are presented in this paper. To our knowledge, this study is the first to demonstrate sex-specific differences in associations between adipokine z-scores and waist circumference z-score and lipids, respectively in healthy children and adolescents.

Unaffected bone mineral density in Danish children and adolescents with type 1 diabetes.

AIMS: Adults with type 1 diabetes mellitus (T1D) have decreased bone mineral density (BMD). Our study aimed at determining BMD and the association to metabolic control in children and adolescents with T1D. METHODS: 244 patients (113 girls) with a median age of 14.3 years and T1D duration of 1-16 years were included. A dual-energy X-ray absorptiometry scan assessed BMD Z-scores excluding the head (total body less head, TBLH). TBLH-BMD were then investigated for associations to diabetes relevant variables such as HbA1c, insulin treatment, anthropometry and physical activity. RESULTS: In all participants the TBLH-BMD Z-score (0.22 ± 0.96) was significantly higher than the references. Separated by sex, TBLH-BMD Z-score in boys (0.11 ± 0.84) was no different from healthy peers whereas TBLH-BMD Z-score was significantly higher in girls (0.36 ± 1.09). The higher TBLH-BMD Z-score in girls were explained by higher BMI Z-scores. Participants with assumed final height (based on age) had an average TBLH-BMD Z-score of 0.78 ± 1.06, significantly higher than references independent of gender, HbA1c, height- and weight Z-scores. Multiple regression analyses showed that TBLH BMD Z-score associated negatively to HbA1c (P = 0.003), pump treatment (P = 0.019) and screen-time (P = 0.005) and positively to weight Z-score (P < 0.001). Physical activity, sex and puberty did not significantly associate to TBLH-BMD Z-score. CONCLUSION: Unlike adults with T1D, BMD is not decreased in children and adolescents with T1D and even elevated after attained final height. As HbA1c negatively associates to BMD, decreased BMD may progress over time. Whether changes in microarchitecture or bone metabolism precede changes in BMD needs further investigation.

High prevalence of disordered eating behavior in Danish children and adolescents with type 1 diabetes.

OBJECTIVE: A higher prevalence of disordered eating behavior (DEB) has been demonstrated in children and adolescents with type 1 diabetes (T1D) compared to healthy aged-matched peers. DEB is associated with higher HbA1c levels and increased risk of developing complications to T1D. The aim of this study was to determine the prevalence of DEB in a Danish cohort of children and adolescents with T1D aged 11 to 19 years and to characterize them regarding metabolic control and relevant clinical data. RESEARCH DESIGN AND METHODS: In a cross-sectional study, we determined the prevalence of DEB using the revised Diabetes Eating Problem Survey (DEPS-R) questionnaire. HbA1c and relevant clinical data were obtained at the time they filled in the questionnaire. RESULTS: Hundred and ninety-two children and adolescents (46% girls) aged 11 to 19 years with T1D were included from the pediatric diabetes outpatient clinic. A total of 40 participants (21%) had DEB. The prevalence was higher among girls compared with boys (34.1% vs 8.9%) and those who had DEB were older (16.7 vs 15.0 years, P < .001), had longer duration of T1D (7.5 vs 4.9 years, P < .001), higher BMI Z-scores (1.2 vs 0.3, P < .001), higher HbA1c (72.8 (8.8%) vs 62.0 (7.8%) mmol/mol, P < .001), higher total cholesterol (4.6 mmol/L vs 4.2 mmol/L, P = .0048), and LDL (2.7 vs 2.3, P = .001) compared with those with no signs of DEB. CONCLUSION: As in other countries, the prevalence of DEB is high in Danish adolescents with T1D. Early detection of DEB is essential to prevent short- and long-term complications to T1D.

Episodes of severe hypoglycemia is associated with a progressive increase in hemoglobin A1c in children and adolescents with type 1 diabetes.

OBJECTIVE: To investigate the trajectory in glycemic control following episodes of severe hypoglycemia (SH) among children and adolescents with type 1 diabetes (T1D). METHODS: A Danish national population-based study comprising data from 2008-17. SH was defined according to the 2014 ISPAD guidelines. A mixed model was applied with HbA1c as outcome and SH episodes and time since first episode as explanatory variables. Data were adjusted for age, gender and diabetes duration. RESULTS: A total of 4244 children (51.6% boys) with 18 793 annual outpatient visits were included. Mean (SD) age at diabetes onset was 9.0 (4.1) years. Median diabetes duration at inclusion in the study was 1.2 (Q1 = 0.9, Q3 = 3.0) years, and median diabetes duration at last visit was 5.0 (Q1 = 2.7, Q3 = 8.1) years. A total of 506 children experienced at least one episode of SH during the nine-year follow-up; 294 children experienced one episode, 115 two episodes and 97 three or more episodes of SH. HbA1c increased with episodes of SH and in the years following the first episode. The glycemic trajectory peaked 2 to 3 years after an SH episode. The accumulated deterioration in glycemic control was in the range of 5% in patients with two or more episodes equivalent to an increase in HbA1c of 4 mmol/mol (HbA1c ~0.4%). CONCLUSION: SH was followed by a progressive and lasting increase in HbA1c among Danish children and adolescents with T1D. Thus, in addition to the known risk of new episodes of hypoglycemia and cognitive impairment, SH contributes to long-term diabetes complications.

Decreased markers of bone turnover in children and adolescents with type 1 diabetes.

BACKGROUND AND AIM: Adults with type 1 diabetes (T1D) have increased risk of bone fractures and decreased bone mineral density (BMD). Alterations in bone turnover have been suggested as the link between T1D and the impaired bone health. Furthermore, bone turnover has been suggested to have beneficial effects on glucose metabolism. This study aimed at describing bone turnover markers (BTM), and the relationship with glycemic control, in children and adolescents with T1D. METHODS: A total of 173 (47% girls) children and adolescents aged 7.7 to 17.5 years with T1D for more than 1 year were included. Participants were evaluated by BMD together with measurements of selected BTM; two formation markers: osteocalcin (OCN) and procollagen type-1 amino-terminal propeptide (P1NP) and one resorption marker, C-terminal cross-linked telopeptide of type-1 collagen (CTX). BTM were converted into Z-scores utilizing new national references. RESULTS: Mean OCN Z-score (-0.68 ± 1.31), P1NP Z-score (-0.33 ± 1.03) and CTX Z-score (-0.43 ± 1.10) were all significantly lower than the reference population (P < .001). No associations were seen between BTM and T1D duration. BMD Z-score was comparable to the reference population and associated with none of individual BTMs. CTX Z-score was negatively associated with HbA1c (P = .007) independent of both exogenous and residual endogenous insulin. CONCLUSIONS: Markers of bone formation and resorption were decreased in children and adolescents with T1D. CTX Z-score associated negatively with HbA1c adjusted for insulin treatment and endogenous insulin production indicating a potential association between CTX and insulin sensitivity. The long-term consequences of decreased BTM on BMD need further attention.

Screening for retinopathy in children with type 1 diabetes in Denmark.

BACKGROUND/OBJECTIVE: Children with type 1 diabetes (T1D) are screened regularly for retinopathy with fundus photography to prevent visual impairment. According to Danish national guidelines, screening should take place at age 12, 15, and 18 years after minimum 3 years of diabetes. As glycemic control has improved, prevalence of retinopathy is expected to be decreased. The aim of this study is to investigate the prevalence, degree, and progression of retinopathy in children with T1D and to explore if screening at 12 years is currently indicated in Denmark. METHODS: Data on all Danish children with onset of T1D from 2003 to 2013 (n = 2943) were collected from the "DanDiabKids" registry. For children with registered screenings (n = 2382), prevalence of retinopathy at 12, 15, and 18 years was determined. In children with retinopathy, subsequent screenings were studied to reveal if retinopathy was persistent or temporary. RESULTS: Prevalence of retinopathy at 12, 15, and 18 years was 0.9%, 2.3%, and 3.1%, respectively. Minimal background retinopathy was seen in over 90% and 100% at 12 years. In available re-screenings, retinopathy resolved spontaneously in 87.5% of all cases and 100% of cases at 12 years. CONCLUSIONS: The prevalence of retinopathy in Danish children with T1D was low. At 12 years, prevalence was 0.9% and exclusively minimal background retinopathy with 100% remission in re-screenings. Thus, screening at this age does not seem to have significant clinical relevance. We propose more individualized screening selection before the age of 15.