Use of Insulin Pump in Neonates and Toddlers

Diabetes management for very young children is challenging and little reported, especially for those younger than 1 or 2 years of age. The report from the DPV in this issue of the journal ¹ will be helpful in initiating continuous subcutaneous insulin infusion (CSII) in these young children. There have been reports over the past 15 years that insulin pump therapy is superior to injection therapy in young children, those younger than 5–6 years of age. Although some reports do not note a meaningful improvement in hemoglobin A1c (HbA1c), reports in children with higher initial HbA1c levels show a decrease in HbA1c after 3 months with sustained improvements. ^{2 –5} The advantage of CSII therapy is thought to be due to the smaller doses (0.025 units) that can be delivered via CSII as a bolus dose or as an hourly basal rate, avoiding larger doses and subsequent hypoglycemia. Likewise, suspending even these small doses may be helpful in correcting hypoglycemia or impending hypoglycemia. This view is supported by a report in the October issue of Diabetes Technology & Therapeutics of improved glycemia without increased hypoglycemia when insulin administered by CSII was diluted to U10 for use in very young children (1–4 years old), allowing even smaller doses to be accurately administered. ⁶ However, further work is needed to define any risks that may be associated with using larger amounts of diluent and if confusion occurs in dosing when using U10 insulin as noted by Frohnert and Alonso. ⁷

Because the number of young children treated at any one center is small, reports of children with type 1 diabetes focusing on those less than 1 or 2 years of age are not available to guide therapy. Virtually all information on insulin treatment, especially insulin pump therapy in children younger than a year of age, is provided by isolated case reports and small series, mainly of infants with neonatal diabetes, treated with CSII. The small number of very young children is illustrated by those treated at the Barbara Davis Center (Aurora, CO), where we care for 3,600 pediatric patients, but only 2.5% were diagnosed prior to a year of age. The T1D Exchange Clinic Registry with >25,000 participants ranging in age from infants up to >90 years of age has reported only 3.2% of participants diagnosed at <5 years of age. ⁸ The DPV registry includes 18,709 children and youth <20 years of age treated in 232 centers from January 1995 and September 2013 and is large enough to help us better define the treatment of children with diabetes diagnosed before 1 year of age. Even within this registry only 344 were <1 year at the time of assessment, or 1.84% of the children in the registry. Of these, 169 (48.8%) were treated with CSII. Therefore, guidelines from the experience for initiating insulin pump therapy in this relatively large cohort of 169 very young children are very helpful. ¹

The report includes only those very young children treated with CSII. It is of interest that 67 of the children included in the CSII study were diagnosed with diabetes prior to 6 months of age so most likely have monogenic neonatal diabetes mellitus (NDM). The remaining 101 children were diagnosed between ≥6 and 12 months of age. Although antibody determinations were not available for all of those ≥6–12 months of age at diagnosis, those with positive diabetes-associated autoantibodies (the type 1 diabetes mellitus [T1DM] group; n = 58) had a similar age of onset to those without antibodies or for whom antibodies were not done (autoantibody unknown diabetes mellitus [the AUDM group]; n = 43) and required similar insulin doses; therefore, all or most are thought to have autoimmune diabetes. The report suggests that glycemic control in these very young children treated under usual care outside of a study was overall good for such young children. HbA1c met International Society of Pediatric and Adolescent Diabetes (ISPAD) guidelines in the NDM group (6.75%) despite two episodes of ketoacidosis in this group and was close to, although above, the ISPAD goal of <7.5% in the two other groups (7.6% and 7.8%).

There was a significant difference in the insulin needs of those with insulin-treated NDM and of those with T1DM. The children with T1DM required a similar total daily dose of insulin (0.60 U/kg/day) but more prandial insulin (0.30 U/kg/day) compared with those with NDM (0.52 and 0.09 U/kg/day, respectively). Of these CSII users, the insulin dosing for those younger than a year of age was very similar to the dosing (in IU/kg/day) found in children 1–5 years of age in the DPV registry. Likewise, the daily insulin profile or hourly basal insulin dosing was similar to that of children 1–5 years of age.

The DPV authors ¹ nicely note that the difference in dosing between the NDM group and the two other groups might be explained because those with NDM were much younger at baseline and taking in smaller carbohydrate amounts more frequently as breast milk or formula. Therefore the prandial doses may have been lower, and the basal rate somewhat higher, as the basal rate was likely providing insulin, at least in part, for the feedings.

It is reassuring that insulin dosing reported in the DPV cohort is similar to that reported by the few others who have reported the insulin dosing in very young children. Alemzadeh et al. ⁹ evaluated children 2–5 years of age (mean, 3.9 years) from initiation of CSII through 12 months of therapy. They used this study to derive a formula for total daily dose (TDD) and for basal dose as a percentage of TDD, as well as a recommended correction factor calculation and an insulin:carbohydrate ratio. The formula was based on continuous glucose monitoring data collected over the 3 months prior to determination of HbA1c, body mass index, and insulin dosing at the visit, with analysis performed at baseline and then every 3 months for a year. Using their formula they found the recommended TDD for young children was 0.74 U/kg/day, and the basal dose was 28% of the TDD. They reported decreased glycemic excursions on CSII compared with prior multiple daily injection (MDI) therapy. In a randomized, controlled trial of CSII versus MDI in children 1–6 years old, Fox et al. ² reported a starting CSII dosing as a TDD of 0.6 IU/kg/day. HbA1c levels at baseline and 6 months were not different between groups, and both met the ISPAD goal of <7.5% (7.24% in the CSII group vs. 7.46% in the MDI group at 6 months).

Because this report also includes a large series of infants with NDM, it is important to comment on the care of these children as well as the experience in CSII dosing in these infants. The DPV registry has reported 90 children with NDM (diabetes diagnosed prior to 6 months of age), and 30 of these had a genetic diagnosis with 17, or approximately 50%, with KCNJ11 or ABCC8 mutations responsive to sulfonylurea. ¹⁰ In the CSII article ¹ the authors note that the 67 NDM patients included who were treated with CSII were not started on CSII until 32 days after diagnosis compared with less than a week after diagnosis, often on the day of diagnosis, in infants diagnosed after 6 months and presumed to have T1DM. The authors suggest the longer time lapse between diagnosis of diabetes and initiation of CSII in those with NDM may be explained by the delay required to confirm the genetic diagnosis, which they suggested might take several weeks.

Reports have suggested that for sulfonylurea-responsive KCNJ11 or ABCC8 mutations, speed in diagnosis is of the essence. Identifying these children and initiating sulfonylurea therapy quickly may have important long-term effects on neuromuscular and perhaps neurocognitive function. ^{11 –14} Therefore, prompt genetic determination and initiation of sulfonylurea therapy may be critical. In some locations, insulin pump therapy may be possible while awaiting genetic testing, avoiding prolonged intravenous insulin therapy and especially avoiding injection insulin therapy for all of the reasons described above. A delay of a month for genetic determination prior to initiation of sulfonylurea should not be necessary. ^14,15 Testing for Kir6.2 and SUR1 mutations in a newborn or young baby should be considered an urgent priority by the referral genetic laboratory. Several genetic laboratories have grant funding to provide testing for Kir6.2 and SUR1 mutations, making testing available almost anywhere in the world. Recent publications do suggest both more rapid testing is being performed in NDM ¹³ and also more efficient testing processes ¹² are being used.

As we learn more about effective insulin therapy in infants and very young children, we should not forget the neonates who will greatly benefit from noninsulin therapy.