Abstract
In this issue of Diabetes Technology and Therapeutics, Frontino et al. 5 describe their experiences with sensor-augmented pump therapy in young children who attended one of three pediatric diabetes clinics in Italy. The authors identified 28 children 7 years of age or younger, already on insulin pump therapy for at least 6 months, who initiated continuous glucose monitoring (with the Dexcom [San Diego, CA] Seven Plus™ sensor), and followed their progress on CGM for an average of 9 months. Overall, they noted very high levels of compliance with sensor wear and generally high reports of parental satisfaction (based on a 1–3 score rating). Specifically, parents reported greatest benefit on reducing fear of hypoglycemia and usefulness of alarms. Glycemic control was on the whole unchanged after the introduction of CGM; although the authors report that the subgroup with the highest baseline A1C did experience a statistically significant 0.9% drop in A1C by the end of the study, it is unclear whether this was a preplanned analysis or how many subjects this subgroup comprised.
The most impressive feature of this retrospective observational study is the high CGM use among these young patients. Although parents did report the usual barriers to compliance (skin reactions, sensor malfunctions, and conflicts with children over use), subjects were still wearing the sensor on average 83% of the time, with 64% of subjects wearing the sensor on a continuous basis. Although almost one-third of the group discontinued the use of the sensor at the end of the study, it appears that in at least several instances the major motivation for discontinuation was economic rather than clinical. These usage data are consistent with a recent Diabetes Research in Children Network (DirecNet) randomized clinical trial of CGM in young children 4–9 years of age, in which average sensor use decreased only from 106 h/week at the beginning to the study to 85 h/week at 6 months, with 41% of subjects averaging at least 6 days/week of wear in Month 6. 6 Similarly, in a feasibility study conducted by the DirecNet group in 20 very young children <4 years of age, 50% of subjects were still using the sensor at least 6 days/week over 6 months of observation. 7
Even given the slight deterioration of sensor use over the course of these 6-month studies, most (if not at least half) of subjects were using the sensor for a majority of the time. Parents were highly satisfied with the devices, with Frontino et al. 5 reporting a reduction in fear of hypoglycemia with the use of CGM, as well as usefulness in daily diabetes management. The DirecNet studies as well demonstrated a high degree of parental satisfaction with CGM: in the study involving 4–9 year olds, the average item rating score was 3.9 out of a 5-point Likert scale, and 86% of scores were ≥3.5. Parents of the toddlers <4 years of age were even more enamored of the device, with 100% reporting favorable impressions of CGM and giving an average item score of 4.1 out of 5. Although fear of hypoglycemia scores were not significantly lower in the CGM group compared with the control group in the DirecNet randomized clinical trial in 4–9 year olds, over 90% of parents reported that CGM use made adjusting insulin doses easier, enabled them to identify patterns not previously visible to them, and made them feel safer about hypoglycemia, themes that were not only echoed but magnified among parents in the toddler sensor study.
If these studies suggest such a love-fest between parents of younger children with type 1 diabetes and CGM, several nagging questions remain, most prominently why wasn't CGM use associated with improvements in glycemic control or reduction in hypoglycemia exposure in these studies? Adult subjects in the JDRF CGM study experienced an average 0.5% drop in A1C, and even in the child and young adult subgroups of the JDRF CGM study, in whom consistent sensor use was much less common, subjects using CGM at least 6 days/week experienced significant decreases in A1C levels over time, as much as 0.5–0.7%. In all three of the CGM studies targeted to young children, despite high levels of sensor wear, overall control and hypoglycemia were unchanged, and neither A1C improvement nor hypoglycemia reduction correlated with amount of CGM use.
The answer to this question lies almost certainly in the reality that despite the well-known rationale for tight control and the growing body of evidence that hyperglycemia may also be detrimental to the developing brain, 8,9 the primary motivation in the management of young children with type 1 diabetes is avoidance of hypoglycemia and, by extension, the mitigation of anxiety related to the risk of hypoglycemia in a young child. This motivation can be found not only among parents of young children with diabetes, but in their clinicians as well, as evidenced by the persistent age-stratified A1C targets recommended by the American Diabetes Association. 10 In the hands of parents and clinicians whose primary goal is to avoid hypoglycemia, overall glycemic control may not necessarily improve, and indeed average glucose levels may rise, if they are successful in limiting hypoglycemia exposure. Ironically, however, the results of the studies to date do not show that parents and clinicians have been effective in avoiding hypoglycemia in their young children: hypoglycemia rates in the DirecNet study of 4–9 year olds were similar between the two treatment groups, 6 and the toddler observational studies of both Frontino et al. 5 and DirecNet 7 show similar hypoglycemia rates before and after the introduction of CGM. And, while exposure to hypoglycemia has not been effectively reduced, parents still report great comfort in the knowledge that they are indeed helping to protect their children.
It is clear that sensor use in a fully open-loop format is woefully inadequate to control blood glucose levels effectively in the majority of pediatric patients. Despite the admirable vigilance of parents who manage the extra inconveniences of using a sensor, endure the complaints of their child, and tolerate the interruptions of sleep inevitably associated with alarms, extra blood tests, and glasses of orange juice, we will only see meaningful reduction in hypoglycemia exposure in this population with closed-loop control. It will require, at the very least, automatic pump suspension, but, more likely, predictive pump suspension algorithms or more advanced fully integrated systems, to achieve our combined and sometimes conflicting goals of improved A1C and avoidance of hypoglycemia.