Automated Insulin Delivery for People With Type 1 Diabetes and High HbA_1c: Is It Time?

It is now more than 100 years since the discovery of insulin and 14-year-old Leonard Thompson becoming the first person, and the first youth, to receive lifesaving insulin injection therapy (1). While this was a revolutionary discovery, Leonard subsequently died prematurely aged 27 years (1). Fast-forward to the 21st century, and while quality of life and traditional therapy have considerably improved for people with type 1 diabetes, reductions in life expectancy of up to 17 years are still reported for young people diagnosed before the age of 20 years in Europe (2,3). It therefore remains clear that children, youth, and adults living with type 1 diabetes continue to experience excess disease and treatment burden, and the majority still do not meet the glycemic targets known to reduce complications and help prevent premature death (4–6). These data highlight that we cannot be complacent and new therapeutic solutions are urgently required.

Advanced diabetes technology, and specifically automated insulin delivery (AID), holds much promise to help with this serious situation, and AID is now recognized as the gold standard therapy for people living with type 1 diabetes (7). However, in examination of the pivotal trials of the different commercial AID systems, one finds that participants are often enthusiastic early adopters and recruited from specialized diabetes services. Mean baseline HbA_1c (±SD) of these participants spans from 7.2% ± 0.9% to 8.0% ± 0.6% (55 ± 9.4 to 63 ± 7.0 mmol/mol) (8–11). These data are not necessarily representative of outcomes from the afore-mentioned clinical databases, where HbA_1c outcomes in adults sit between 7.7% ± 1.2% and 8.9% ± 1.9% (61 ± 13.1 and 74 ± 20.8 mmol/mol), particularly for adolescents, where HbA_1c outcomes sit above 9.2% ± 1.9% (77 ± 20.8 mmol/mol) (5). Interestingly, in one of the first AID studies published in Diabetes Care, a subanalysis of youth compared with the whole sample demonstrated that youth had the greatest HbA_1c improvements (8). This hinted for the first time that those with less-in-target diabetes outcomes may have the most to gain from AID. However, there remains concern that use of insulin pumps and technology may in fact be too burdensome and potentially dangerous for those with very elevated HbA_1c (12,13). There is clearly a need for further evidence on the use of AID in youth (and others) with out-of-target glycemia. In this issue of Diabetes Care, Abraham et al. (14) investigate this critical issue.

Abraham et al. undertook a randomized controlled trial comparing 6 months of AID with usual therapy (in this case continuous subcutaneous insulin infusion [CSII] therapy with or without continuous glucose monitoring [74% of participants also were also current continuous glucose monitor users]) in 42 youth with mean HbA_1c at baseline of 9.8% (84 mmol/mol). Their main finding was a mean adjusted group difference in HbA_1c of 0.77 percentage points (95% CI −1.45 to −0.09) (8.4 mmol/mol; P = 0.027) favoring AID over usual CSII therapy. This also resulted in an increase in time in target glucose range of 19.1 percentage points in the AID group, with no evidence of increased hypoglycemia risk. Interestingly, despite these improvements in glycemia, there were no differences in psychosocial outcomes between the two groups.

First, the context of this study is important. This is a small study where the recruitment target of 44 participants was not met. Nonetheless, randomized controlled trial level evidence is a considerable strength. In addition, studies including particpants with very elevated baseline HbA_1c, ranging from 8.7% (72 mmol/mol) to 13.7% (126 mmol/mol), are few, and with a mean age of 16.2 years, this study is clearly focused on a population with type 1 diabetes who have considerable and complex needs. The results are something to celebrate but also leave many questions. Improving HbA_1c by 0.77 percentage points is an achievement in this population and a considerable improvement over levels with their traditional baseline CSII. In addition, no adverse safety signals were seen, which provides much-needed encouragement regarding offering AID in this population. However, HbA_1c gains were less than the usual HbA_1c results seen with use of AID in more traditional populations, usually in the order of >1 percentage point (10–14 mmol/mol) (8,11). In a recent single-arm trial with use of an identical AID system in 20 youth with higher mean HbA_1c at baseline (10.5%), a 2.9-percentage-point (31 mmol/mol) improvement in HbA_1c was seen (15) and sustained out to 12 months (16). Baseline therapy differed, though, with direct transition from multiple daily injections—as opposed to CSII in this study. These positive findings have also been replicated in a recently published 80 person RCT (17).

So why have we not seen those gains here, and what do we take from these results? Some clues are provided for us to consider. Firstly, these participants were already struggling not only with glycemia but also with use of technology. They all were existing users of CSII, and despite access to publicly funded CGM systems only 74% were using them at baseline, and with only 43% sensor wear in the prior 3 months. By the end of the study (6 months of AID use), there were no improvements in these outcomes, with mean sensor use time at only 46% and the amount of time in automation very low, at 49%. Clearly, if it is not used, AID cannot improve outcomes. The question of how to support these young people in their use of AID is clearly crucial. Why is it that some thrive with AID and others do not? This is a space where further research is urgently needed, but what is clear is that these young people need access to a high level of support, for training and safety and for provision of motivation and encouragement.

Abraham et al. also note the high baseline levels of diabetes distress: experienced in >70% of the participants. This is a very important observation and highlights the burdens faced by young people living with type 1 diabetes. For many, AID alone may facilitate changes in both glycemia and behavior, leading to improved health outcomes. For others already experiencing high burden and other social challenges, a more complex package of support including education, psychological, and social support components is likely required.

Use of AID by youth with type 1 diabetes is a challenging but also very exciting area of research and clinical practice. Not all youth with complex needs will respond well to AID, but data are now clearly starting to emerge showing that AID use in youth populations with complex needs can further improve glycemia, with no additional safety alerts. Going one step further, we observe that those with high HbA_1c have in fact the most to gain from current, and future, AID technologies and should not be excluded from what has become the new gold standard in care. Further work is clearly required to improve the technology, wear burden, and wraparound support offered. We also need to ensure equity of access to all people with type 1 diabetes, regardless of their background.