Update: Pediatric Diabetes

1.1 New closed-loop device for type 1 diabetes

Technology has revolutionized the management of T1D. The use of continuous glucose monitors (CGM) reduces the frequency of hypoglycemia.1 However, multiple daily injections and insulin pumps are the only means of insulin administration at present. The only Food and Drug Administration approved “artificial pancreas” or closed-loop insulin pump is the Medtronic's MiniMed 670G pump, which is a hybrid system that is able to automatically adjust basal insulin but requires manual input of blood glucose and carbohydrate intake with meals.

In the October 2019 issue of the New England Journal of Medicine, data from the 6-month International Diabetes Closed Loop (iDCL) trial of a new closed-loop system for T1D were published.2 This study, conducted in 7 United States university hospitals as a parallel, unblinded study for 26 weeks compared the safety and efficacy of the Control IQ t:slim X2 closed-loop insulin pump to a sensor-augmented open-loop insulin pump (without low glucose suspension feature). This closed-loop system provides automated basal and bolus insulin delivery.

Patients aged between 14 and 71 years of age with a diagnosis of T1D and on insulin (pump or multiple daily injections) for at least 1 year were included in the trial. They were randomized in a 2:1 ratio to use the closed-loop system (112 participants) vs a continuous glucose monitor (Dexcom G6) augmented pump (58 participants) without a cutoff for hemoglobin A1C (HbA1C, 5.4 to 10.6% in both groups). Patients had a 2-8 week run in period followed by visits at 2, 6, 13, and 26 weeks and multiple telephone encounters in between, maintaining 100% retention at the end of the trial. The primary outcome of this trial was the percentage of time spent in glucose target range (70-180 mg/dL) as measured by the Dexcom G6.

Glucose levels in the target range increased from 61 ± 17% at baseline to 71 ± 12% during the 6 months in the closed-loop group whereas it remained at 59 ± 14% in the control group (95% confidence interval [CI], 9 to 14; P < 0.001). This meant that those in the closed-loop group spent 2.4 hours more time per day within target range and this finding was sustained throughout the study. The greatest difference in the median percentage of time in the target range was at 5 am (89% in the closed-loop group vs 62% in the control group). As a secondary outcome, it was found that the closed-loop group spent a significantly lesser percentage of time in hypoglycemia <70 mg/dL (1.58 ± 1.15) compared to the control group (2.25 ± 1.46) (P < 0.001). They also spent a significantly lesser percentage of time with glucose <54 mg/dL compared to the control group (0.29 ± 0.29 vs 0.35 ± 0.32, P = 0.02). The mean difference in HbA1C at 26 weeks was −0.33 percentage points (95% CI, −0.53 to −0.13; P = 0.001). The primary outcome was evaluated following consideration of age, sex, body-mass index, income, educational level, insulin pump or injection use, previous use of a continuous glucose monitor, and glycated hemoglobin level as covariates and remained consistent across all 7 institutions. Although severe hypoglycemia was not reported in either groups, 12 episodes of hyperglycemia and 1 diabetic ketoacidosis (DKA) occurred in the closed-loop group compared to 2 episodes of hyperglycemia and no DKA in the control group. All these events were attributed to infusion set failures.

Overall, the new t:slim closed-loop system seems to be a promising form of technology that T1D patients could take advantage of once it becomes commercially available. The strength of this study includes its 100% retention rate and the median time spent in closed-loop mode being 90% over 6 months. However, more unscheduled visits and pump failures may require further evaluation. Although HbA1C was not a criterion for selection of patients, the majority of patients in both groups had an average HbA1C of 7.4%, which indicates good control of T1D. All patients may not be suitable for such a system and it will be important to carefully select those who will benefit from this technology.

1.2 Is T1D in children <7 years of age different from those with a later age of onset?

T1D is a multifactorial disease that is associated with genetic predisposition and autoimmunity resulting in beta cell destruction. Children diagnosed with T1D at an age < 7 years are considered to have a more aggressive phenotype with recent data suggesting significant B cell and T cell proliferation compared to lesser B cell infiltration in those diagnosed in their teen years.3 The HLA region on chromosome 6p21 has been associated with familial cases of T1D, primarily class I and II alleles.4

A recent paper titled “Genetic Variants Predisposing Most Strongly to Type 1 Diabetes Diagnosed Under Age 7 Years Lie Near Candidate Genes That Function in the Immune System and in Pancreatic β-Cells” and published in the January 2020 issue of Diabetes Care attempts to identify the genetic changes that differentiate T1D in children <7 years of age compared to those older than 13 years of age.5 This study had 18 485 controls and 3121 subjects with T1D < 7 years of age, 3757 subjects between 7-13 years and 1708 subjects >13 years of age. Most patients were from the United Kingdom although the cohort included those from Asia Pacific, Europe, and the United States as well. Inshaw et al looked at 17 HLA and 55 non-HLA loci to understand the predictors of T1D. They found that HLA DR3-DQ2/DR4-DQ8 diplotype was more abundant in the <7-year age group compared to the >13-year age group whereas the haplotypes DRB1*15:01-DQB1*06:02 and DRB1*07:01-DQB1*03:03, considered to be protective against T1D, were less prevalent in the <7-year age group. This confirms the previously reported HLA class I and II associations with T1D. The protective haplotype identification was a novel finding in this study. Of perhaps more importance is this group's findings of non-HLA loci associations. They found 7 regions that were differentially associated with the <7-year and > 13-year age groups (false discovery rate [FDR] <0.1), involving cathepsin H (CTSH), GLIS family zinc finger 3 (GLIS3), Ikaros family zinc finger 3 (IKZF3), chymotrypsinogen B1 (CTRB1), the third index variant at interleukin 2 receptor a (IL-2RA), thymocyte-expressed molecule involved in selection (THEMIS), and interleukin-10 (IL-10). The CTSH loci continued to show this differential association between the groups after Bonferroni correction (P<0.05). When a similar analysis was performed only in the UK cohort in order to prove that the findings were not due to population heterogeneity, 6 of the 7 loci (except CTRB1) continued to show a differential association between the age groups. Similar associations were identified when <5-year-old and < 6-year-old groups were evaluated indicating this was not purely due to chance. The GLIS3 loci is a transcription regulator that is expressed in the beta cells of the pancreas. The remaining loci have their actions in T cell and/or B cell biology. Hence, in the younger population, the T1D phenotype is associated with HLA class II alleles as well as loci that affect lymphocytes resulting in immune infiltration and beta cell destruction. The CTSH, IKZF3, and THEMIS loci also colocalized with whole-blood expression quantitative trait loci (eQTLs) making this possible causative for T1D risk.

This study brings to light the important interactions of genes, immune proliferation, and lymphocyte action causing beta cell destruction. The identification of some of these “risk loci” may pave a way for further research into the etiology and from there, the cure for T1D. The smaller number of patients in the >13-year group may be a limitation of the study. These findings will need to be replicated.

1.3 Is there a difference in mortality rates in young adults with T1D?

Long-standing diabetes is associated with multiple comorbidities including retinopathy, nephropathy, coronary artery disease, and neuropathy.6 A Swedish study published in Pediatric Diabetes in January 2020 looked to calculate the mortality rates in people <29 years of age with T1D diagnosed <18 years of age and to understand its association with metabolic control of the diabetes.7

Samuelsson et al utilized the SWEDIABKIDS database (SWE) established in Sweden in 2000 with 43 pediatric diabetes centers, covering 99% of children and adolescents with T1D and The Swedish Cause of Death Register (CDR) to understand mortality rates. The study identified 12 652 patients (54% males) diagnosed with T1D before 18 years of age among whom 68 (0.5%) were deceased <29 years of age between 2006 and 2014, 36 males and 32 females. Diabetes mellitus was the cause of death recorded for 26 deceased patients (38%), 6 of whom were below 18 years of age at the time of death. The codes available in the CDR did not differentiate between acute hypoglycemic coma vs diabetic ketoacidosis. In this group of patients, 22 died from acute diabetic coma (hypo- or hyperglycemia). The standardized mortality ratios (SMR) were significantly higher for females who died due to cardiovascular causes (SMR 8.7, 95% CI 2.8-21.0) and neurological causes (SMR 10.9, 95% CI 4.0-24.1). Other causes of death included poisoning in 3, accident in 6, suicide in 6, malignancy in 2, infection in 2, and unknown causes in the others. The SMR for this cohort was 2.7 (2.1-3.4, 95% CI) with male SMR 2.0 (1.4-2.7, 95% CI) and female SMR 4.4 (3.1-6.2, 95% CI). The SMR increased with age with the highest ratio of 6.3 (3.4-10.7) in the 25-29 years age group. When diabetes control was analyzed, those deceased from diabetes causes had a significantly higher (P < 0.001) HbA1C at 8.9 ± 1.7% compared to those alive at 7.8 ± 1.1%. There was no difference in the HbA1C levels between the patients who were alive and those deceased from all other causes. Although the likelihood of death registered as due to diabetes was higher with higher HbA1C levels (hazard ratio 1.06, 95% CI 1.02-1.10), there was no such likelihood noted for other causes of death. Age of diabetes onset was not related to risk of death.

This is a single-nation study that highlights the importance of registries in long-term follow-up of patients with T1D. The authors show that in Sweden, there is an increased risk of premature deaths in young adults with T1D compared to the general population, primarily due to diabetes causes. However, the exact cause of death is not well described because data were collected based on International Classification of Diseases code registered at the time of death, which does not differentiate hypoglycemia from hyperglycemia. A majority of the deaths occurred due to acute causes unlike in older populations where macro and microvascular complications are associated with higher morbidity and mortality. A total of 0.6% of the deaths that were caused by suicide should not be ignored because chronic conditions like T1D have significant impacts on mental health and quality of life. Longer studies involving multiple nations will be helpful in identifying the true risk of premature death associated with T1D.