MiniMed 780G Provides Safe and Effective Short- and Long-Term Management of a Person with Type 1 Diabetes with an Extremely Elevated HbA1c and Multiple Hospital Admissions for Cannabis Intoxication: A Case Report

Abstract

Adolescents and young adults (youth) with type 1 diabetes (T1D) often face a significant diabetes burden, and optimizing glycemia presents a particular challenge. These challenges are compounded by physiological and psychosocial complexities, which can create barriers to accessing advanced diabetes technologies, which are the current gold standard of care. This case report highlights a young individual who, to our knowledge, represents the highest HbA1c level ever published at the initiation of automated insulin delivery. It explores his experiences with automation, safety, and multiple hospitalizations related to intoxication.

Introduction

Type 1 diabetes (T1D) is caused by autoimmune destruction of pancreatic beta cells, leading to lifelong insulin deficiency. Recent studies have indicated a global increase in T1D incidence with a high prevalence among adolescent and young adult populations (youth).^1–4 Youth also experience considerable difficulties with diabetes burden and in achievement of the glycemic goals known to prevent acute and long-term diabetes complications.^5,6 This is due to several factors including physiological changes, social pressures, family pattern, developing independency, and risk-taking behaviors.^3,7,8 One of the challenging problems is use of recreational drugs such as cannabis and alcohol, which further complicates glycemia in this group.^9–14

Advanced diabetes technology, particularly automated insulin delivery (AID) systems, shows significant potential in improving glycemia, enhancing quality of life and reducing diabetes-related complications in youth.^15–21 While AID is now the gold standard therapy for people with T1D, youth and those with above-target diabetes are often considered unsuitable candidates for AID. This is due to safety concerns, especially in those with alcohol, cannabis, and other substance use or who are facing particular challenges with traditional diabetes self-care and treatment adherence.^9,22,23 Here, we present a case of a young adult we believe demonstrates the highest HbA1c ever published to commence AID and his experience with automation, safety, and multiple hospitalizations for intoxication.

Case Presentation

In light of long-standing extremely elevated glycemia (most recent HbA1c of 17.6% [169 mmol/mol]) and recurrent hospitalizations, a 23-year-old male diagnosed with T1D for 13 years and using multiple daily injections (MDI) and capillary glucose measurements presented for consideration of AID as part of a research trial. ²⁴ This person with diabetes had never used any form of insulin pump therapy before participating in the trial. Prior to puberty his HbA1c had fluctuated between 8% and 10% (64–86 mmol/mol) but had subsequently deteriorated ranging from 10% to 17.6% (86–169 mmol/mol). Over this time and despite full multidisciplinary team input (medical, nursing, dietitian, and psychology), he frequently missed clinic appointments and developed multiple diabetes-related complications, including recurrent episodes of diabetes ketoacidosis (DKA), retinopathy, and low mood. His body mass index (BMI) decreased from 26.5 kg/m² to 17.9 kg/m², and he commenced smoking cigarettes, cannabis, and consuming alcohol. His fasting lipid profile also deteriorated during this time (total cholesterol [Chol] 5.4 mmol/L, LDL (low density lipoprotein) Chol 3.0 mmol/L, Chol-to-HDL (high density lipoprotein) Chol ratio 3.1, triglyceride 1.4 mmol/L at 3 years pre-AID; Chol 6.5 mmol/L, LDL Chol 3.5 mmol/L, Chol-to-HDL Chol ratio 4.7, triglyceride 3.6 mmol/L at 1 year pre-AID). There was no evidence of hypertension (122/85 mmHg) nor microalbuminuria (urine albumin-to-creatine ratio <1.0 mg/mmol). He was hospitalized due to multiple episodes of DKA and symptoms mimicking DKA (nine hospitalizations following diagnosis), which were at times due to complications of substance use including alcohol and cannabis. Additionally, he experienced considerable fear of hypoglycemia, which further complicated his glucose management. As a result, he routinely omitted his rapid-acting insulin, administering only one dose of long-acting glargine insulin daily (estimated total daily insulin dose 0.36 units/kg/day).

Data from 2 weeks of blinded continuous glucose monitoring (CGM) directly prior to starting AID showed that 100% of time was spent in the very high range above 250 mg/dL (>13.9 mmol/L), with no time spent in range 70–180 mg/dL (3.9–10.0 mmol/L) (TIR) or below 70 mg/dL (<3.9 mmol/L). On the morning of his pump start (MiniMed^™ 780G with Guardian^™3 Sensors [Medtronic, Northridge, CA]), the individual presented with ketones of 4.0 mmol/L (normal range <0.6 mmol/L) and an elevated blood glucose above 599 mg/dL (33.3 mmol/L). His BMI was 18.1 kg/m² and HbA1c 17.6% (169 mmol/mol) (measured by formal hospital laboratory). With his clear impending DKA, consideration was given to cancelling his pump start, but in light of his background, current HbA1c (suggesting ketosis was common), and willingness on his part to proceed, he received an insulin injection to correct for hyperglycemia and ketosis, and a pragmatic decision to continue with his training and establish him on pump therapy was made. This was uneventful and his ketones resolved over the single day of training. The training protocol has been published previously. ²⁴ In short, use of sensor-augmented pump with predictive low glucose management was established within 1 day. After 72 hours in the community, automation was enabled by activation of the SmartGuard^™ feature. His personalized settings in manual mode were basal rate of 1.2 units/h, insulin-to-carbohydrate ratio of 9 g, and insulin sensitivity factor of 32 mg/dL (1.8 mmol/L) for daytime and 40 mg/dL (2.2 mmol/L) for nighttime, with a SmartGuard target of 120 mg/dL (6.7 mmol/L). To circumvent the individual’s notable fear of hypoglycemia, and at his request, this target was maintained throughout. Brief data review, setting refinements, and further support were provided daily for 7 days, then weekly for 4 weeks, then monthly out to 6 months, and subsequently 2-monthly out to the end of the 12-month trial. All in-person and remote contact included anticipatory guidance on DKA prevention. No additional therapy or psychological interventions occurred.

Glycemic and insulin delivery profiles of the first 7 days of AID system use are displayed in Supplementary Figure S1 and show that improvements occur rapidly. Figure 1A shows glycemic metrics spanning the first 12 months of AID use, with Figure 1B highlighting HbA1c progression in the years prior to and following AID initiation. As shown, after 3 months of AID use, HbA1c decreased dramatically from 17.6% to 7.9% (169 to 63 mmol/mol). As shown in Figure 1B, after 6, 9, and 12 months, HbA1c remained stable. In addition, the individual appeared much healthier in appearance and demeanor, and his BMI increased to 20.3 kg/m². Furthermore, of the mean total daily insulin dose of 42.5 units, 98.4% was administered by automation (autocorrections and auto basal) and only 1.6% was delivered as a user-initiated bolus, thus demonstrating improved glycemic outcomes despite largely absent meal announcement. He has now used the 780G AID system for three years (transitioning to Guardian 4 sensors after the first 3-months of use), and over this time he has never experienced DKA or severe hypoglycemia.

FIG. 1.

Glycemia prior to and following AID initiation. (A) HbA1c and standard glycemic CGM measures assessed for 2 weeks at baseline and after 3, 6, 9, and 12 months of AID use. (B) HbA1c progression following type 1 diagnosis at age 10 years during both MDI and AID use. AID, automated insulin delivery; CGM, continuous glucose monitoring; CV, coefficient of variation; MDI, multiple daily injections, SG, sensor glucose.

The individual further experienced several psychosocial improvements that he reported. Selected direct quotes are as follows:

When it was at 12 [mmol/L; 216 mg/dL], I would feel like it was low […] My thoughts about the change is [sic] it’s good because I now don’t have to worry about getting low. Like, say if it’s 12, I’m not worried about feeling low or freaking out because it’s down at that […]. I can trust that if it’s around there and if it goes a bit lower, that’s fine.

I don’t have the worry of going low or going too high, so I can go out. Do a lot more than I used to, without that worry. It’s made a dramatic change in my life and what I can do and what I feel comfortable doing.

He also reported on improved sleep: “I sleep through the whole night now. Like before the study, […] I used to get up quite a bit during the night and go to the toilet or I’d just wake up and not go back to sleep and struggle to even go to sleep. But now I fall asleep relatively fast and sleep through the whole night. It’s great.”

However, he was admitted to hospital on two occasions in the first 6 months of AID for hyperemesis related to cannabis intoxication. At time of the first hyperemesis admission, his glucose was 122 mg/dL (6.8 mmol/L) and ketones 1.4 mmol/L. His insulin pump was temporarily discontinued, and he was switched to IV insulin infusion during the critical phase of his condition by the clinical team. Use of AID was safely resumed after recovery. At time of the second admission, he presented with frequent vomiting accompanied by confusion and reduced level of consciousness. Venous blood gas pH was 7.32 (reference range: 7.3–7.43), bicarbonate 30.4 mmol/L (reference range: 22–30 mmol/L), and glucose 348 mg/dL (19.3 mmol/L) with ketones of 2.3 mmol/L. Cannabis hyperemesis was considered the likely cause by the emergency physician, and he was given antiemesis therapy with ondansetron and IV 5% dextrose with 0.9% saline. Normally, his insulin infusion set would have been preemptively changed given the elevated glucose and ketones, but there was a 2-h delay before diabetes team review, at which stage his ketones were already rapidly resolving as was his hyperglycemia, with sensor glucose of 270 mg/dL (15 mmol/L) on CGM at 2 h post-IV fluid. He remained off food, on IV fluids, and with reduced level of consciousness for the following 24 h, and then with reduced appetite and eating for a further 24 h. His CGM and AID trace during the hospitalization is shown in Figure 2. Time-in-range while unwell remained between 93% and 100%.

FIG. 2.

CGM and insulin delivery data during 5 days of hospitalization due to cannabis intoxication and hyperemesis. CGM, continuous glucose monitoring; TDD, total daily insulin dose.

The individual has had no further hospital admissions for any reason in the 2 years following the last episode. At his last clinic visit, 3 years from commencing AID, his weight was 73 kg and BMI was 26.4 kg/m². His latest glycemic data showed 58% TIR and 1% time below 70 mg/dL (3.9 mmol/L), with an HbA1c of 7.6% (60 mmol/mol). His lipid profile has also improved with Chol 4.9 mmol/L, LDL Chol 3.1 mmol/L, Chol-to-HDL Chol ratio 3.6, and triglyceride 0.9 mmol/L. He now has a stable job, and his first child is 1 year old.

Retinopathy grading 10 months prior to AID initiation was R4M3 (severe nonproliferative diabetic retinopathy in both eyes) using a modified “Airlie House Classification Scheme of the Early Treatment Diabetic Retinopathy Study,” ²⁵ which is similar to the “International Clinical Disease Severity Scale.” ²⁶ Subsequent assessments and retinopathy scoring are as follows: 2 months post-AID R2M0 (mild retinopathy); 6 months post-AID R3M0 (moderate retinopathy); 18 months post-AID Left R3M2 (moderate retinopathy), Right R4M2 (severe nonproliferative diabetic retinopathy). While on close observation, he has required no intervention as yet for his retinopathy.

Discussion

This case shows the substantial potential of AID to safely improve glycemia and quality of life in even the most challenging situations. Highlighting this potential is important, as often those with challenging circumstances and/or extremely elevated glycemia have not been considered appropriate candidates to successfully and safely transition into older forms of diabetes technology.

To our knowledge, an HbA1c of 17.6% (169 mmol/mol) is the highest described in the diabetes technology literature. Without doubt, care and caution are required when using technology in this population, but these data clearly show the life-changing potential of AID and related support and care. Few other therapies in diabetes can result in HbA1c improvements of this magnitude, in this case an absolute improvement of 9.7 percentage-points (82.5 mmol/mol). Increasing evidence of efficacy in youth with very elevated glycemia is emerging now. One single-arm study and two recent randomized controlled trials (RCTs) have investigated AID in this population.^24,27–29 These have shown that AID offers significant benefits in improving glycemic outcomes, promoting behavior change, and subsequent improvements in psychosocial outcomes. Nevertheless, comprehensive management of diabetes distress is crucial for effective diabetes management.

While AID has enormous potential to improve glycemia in youth, it is not without risk. However, it is important to understand that there is considerable risk in taking no action in individuals like shown here. For this individual, prior to AID, hospitalization and DKA were frequent and life-threatening occurrences. This is demonstrated in the literature where elevated glycemia and DKA are shown to increase mortality rates and contribute to long-term complications including cognitive impairment and cardiovascular diseases.^30–33 Safely commencing AID in any person with diabetes requires regular anticipatory guidance on DKA and infusion set failure, ³⁴ both commonly described in the literature. ³⁵ As demonstrated here, this can be safely achieved with DKA admissions reducing and overall safety improving. This is also seen in RCT and single-arm study data.^24,27–29 Early worsening of diabetic retinopathy is another significant safety concern given the magnitude of baseline HbA1c and subsequent rapid improvement.^36,37 Ultimately, inaction will lead to significant retinopathy in the long-term as well. While this is a single case report, these data are reassuring in that retinopathy status did not deteriorate. Overall, retinopathy improved for the first 6 months of AID use and was similar to baseline by 18 months. This case is consistent with the relatively limited literature on AID and retinopathy. ³⁸ Importantly, close retinopathy monitoring and, if required, early intervention are crucial for future similar situations. ³⁸

In this case example, while diabetes-related safety improved with DKA admissions ceasing, admissions related to risk-taking behavior and substance use did not. The life period spanning adolescence and young adulthood is complex. It is a time of considerable physical, cognitive, emotional, and psychosocial development. With this comes an increased rate of risk-taking behavior. ³⁹ Young people with chronic diseases including T1D show similar or even increased patterns of these behaviors, which can complicate their outcomes and increase the risk of acute and long-term complications.^5,6,40 Moreover, in youth with T1D, psychosocial factors are implicated in mortality from acute events, which comprise not only acute diabetes-related conditions but also accidents and self-harm. ⁴¹ Onset and use of substances such as cannabis and alcohol peak during this life period, which can further impair safe diabetes self-management. ⁴² In these circumstances, as long as AID is maintained, automated therapy may provide a safety net by minimizing risk of hypoglycemia as well as hyperglycemia, as impressively highlighted in this case in Figure 2. This has also been shown in an earlier case report of severe alcohol intoxication using the same AID system where the automated basal, bolus adjustment, and low glucose alert functions of the SmartGuard feature provided considerable protection during an episode of severe hypoglycemia. ⁴³

In conclusion, this case report highlights the potential of AID to improve glycemia and reduce diabetes burden in a young person with T1D and challenging behavioral aspects. Complex backgrounds should not preclude the use of modern AID. With readiness for change, support, and anticipatory guidance, life-changing outcomes are possible. We hope this report can inspire optimism for both young people with diabetes and the health professionals caring for them.