Abstract
Limited insulin pump cartridge volumes can present challenges to automated insulin delivery system (AID) use for adolescents and young adults (AYA) with type 1 diabetes (T1D) and high insulin requirements. We assessed the real-world safety and effectiveness of U200 concentrated insulin use in AID (U200-AID) among AYAs with T1D. We conducted a two-center, retrospective cohort study assessing glycemia, pump utilization, and safety outcomes pre-/post-U200-AID. Among 50 AYA initiating U200-AID (age 15.4 years, T1D duration 5.5 years, HbA1c 8.5%), time in range (70–180mg/dL) increased (44.6±12.6% vs. 48.9±11.4%, p=0.012) and time below range (<70mg/dL) did not change significantly. Days between cartridge changes increased (2.2±0.5 vs. 3.0±0.5 days, p<0.001) despite increased total daily insulin dose (102.6±23.5 vs. 125.8±38.9 U100 insulin units, p<0.001). No severe hypoglycemia or diabetic ketoacidosis occurred (median follow-up 290 days [IQR 227,476]). These data suggest that U200-AID is a viable option for individuals with T1D and high insulin requirements.
Keywords: type 1 diabetes, automated insulin delivery, concentrated insulin, U200, insulin resistance, insulin pumps
Introduction
The limited capacity of insulin pump cartridges used in automated insulin delivery systems (AID) can act as a barrier to technology use for individuals with type 1 diabetes (T1D) and high insulin requirements(1). U200 concentrated rapid-acting insulin (U200), provides the same dose of U100 insulin (U100) in half of the volume and could potentially facilitate AID use for these individuals(2, 3). To date, evidence for using U200 in AID (U200-AID) has been limited to case reports(3). Despite limited evidence, off-label use of U200-AID has been increasing in real-world settings. The objective of this study was to assess real-world changes in glycemia, pump utilization, and serious safety events in adolescents and young adults (AYA) with T1D who used U200-AID.
Materials and Methods
This retrospective cohort study conducted at the Children’s Hospital of Philadelphia and Johns Hopkins University used a pre-/post study design. The protocol was approved by institutional review boards at both institutes, in accordance with the Declaration of Helsinki and granted a waiver of consent, with data sharing agreements in place for deidentified data.
Study population
AYAs <22 years with T1D initiating U200-AID between January 1st 2020 and September 1st 2024 were identified from the electronic health records (EHR) using ICD-10-CM codes (E10*) and confirmed by chart review. AYAs without verifiable CGM or AID data and those with <16% (14 days) CGM active time in the baseline or follow-up periods were excluded from CGM and pump utilization outcomes (4).
Data Collection
Demographic, clinical, and laboratory data were extracted from the EHR. CGM and AID data were extracted from Dexcom Clarity (baseline CGM data), Tandem Source, or Glooko cloud-based portals. U200-AID start date was determined by EHR documentation and confirmed by insulin dose adjustments in Tandem Source or Glooko.
Outcomes
The primary outcome was change in CGM time in range (70–180 mg/dL, TIR) from baseline to follow-up. Secondary outcomes included changes in time above (TAR, >180 mg/dL) and below range (TBR, Zone 1 [54–69 mg/dL], Zone 2 [<54 mg/dL]), average glucose, coefficient of variation, pump utilization metrics, and safety events. Safety events included severe hypoglycemia, defined as hypoglycemia requiring an emergency room visit or hospitalization, and diabetic ketoacidosis (DKA).
For CGM and pump utilization metrics, baseline was defined as the 90 days pre-U200-AID initiation and follow up as the 90 days post-U200-AID (Supplemental Figure S1). For HbA1c and anthropometrics, baseline was the measurement closest to U200-AID initiation date that fell within 90 days pre- and 14 days post-U200-AID initiation, and follow-up was the measurement closest to 90 days post-U200-AID date that fell within 60–180 days post-U200-AID. Baseline safety events were assessed in the 365 days before and after U200-AID initiation. (minimum 90 days post-U200-AID or until end of study 12/1/2024).
Statistical analysis
Data were assessed for normality using the Shapiro-Wilk test. Continuous variables were reported as mean and standard deviation (SD) or median and interquartile range (IQR), as appropriate. Glycemic metrics, pump utilization data, and safety outcomes were compared pre-/post-U200-AID initiation using paired t-tests or Wilcoxon signed-rank test. Secondary analysis stratified by baseline glycemia (HbA1c <8% vs. ≥8%) was performed. Sensitivity analysis including only subjects using U100-AID before U200-AID initiation was performed to address possible confounding by AID initiation in those who were previously using multiple daily injections (MDI). Complete case analysis was done for missing data. Analyses were generated using STATA software v.18.0 (College Station, TX: StataCorp, LLC).
Results
Fifty AYA meeting inclusion criteria (Supplemental Figure S2) were included. At U200-AID initiation, median age was 15.4 years [IQR 13.9,16.7], 36% (n=18) were female, 68% (n=34) identified as non-Hispanic white, 12% (n=6) as non-Hispanic black, 14% as Latine (n=7), 6% (n=3) as other, and 72% (n=36) were privately insured. Median T1D duration was 5.5 years [IQR 3.0,8.5], median HbA1c was 8.5% [IQR 7.3,9.2], 54% (n=27) had baseline HbA1c > 8.0%, and all subjects were overweight or obese (Supplemental Table 1). Pre-U200-AID, 82% [n=41] used U100AID, 10% [n=5] used MDI with U100 and 8% [n=4] used non-automated pumps. The majority initiated U200-AID with the Insulet Omnipod 5 (OP5) (60%) and the remainder with Tandem t:slim X2 with Control-IQ (CIQ) (40%). One AYA changed from U100 in CIQ to U200 in OP5. Three AYAs discontinued U200-AID within 30 days due to painful infusion sites (n=1), changing to an AID-incompatible CGM (n=1), and insurance-related challenges obtaining pump supplies (n=1).
Glycemic Outcomes
CGM metrics (n=43) are shown in Table 1. Seven AYAs were excluded due to insufficient CGM data pre-U200-AID (n=3), post-U200-AID (n=2) and discontinuing U200-AID within 14 days (n=2). Median CGM active time was >90% (>81 days) and did not differ pre-/post-U200-AID. Mean TIR increased 4.3%-points in the entire cohort from 44.6±12.6% to 48.9±11.4% (p=0.012) and TIR increased ≥ 5% for 52% of the sample (n=22/43). Mean TAR >180mg/dL decreased from 54.2±12.3% to 49.1±13.2% (p=0.003). TBR remained <4%, though there was a statistically significant increase in Zone 1 TBR (median 0.7% [IQR 0.2,1.0] vs. 0.7% [IQR 0.4,1.3], p=0.015) but not Zone 2 TBR. Average glucose decreased from 207±27 to 194.1±27 mg/dL (p<0.001). HbA1c values were missing for three participants at baseline and six at follow-up; for those who had pre-/postU200-AID HbA1c values (n=34), mean HbA1c did not change significantly (8.2±1.1 vs. 8.0±1.2, p=0.07). Median weight increased from 90.0 kg [IQR 78.1,106.8] to 98.8 kg [IQR 83.1,111.2] p<0.001 and BMI increased from 32.3 kg/m2 [IQR 29.2,34.4) to 33.4 kg/m2 [30.4,36.3] p<0.001). In sensitivity analysis including only those using U100-AID before initiating U200-AID (n=37), TIR improved by 3.0%, from 46.3±12.2% to 49.3±10.7% (p=0.082), but did not reach statistical significance and TIR increased ≥ 5% for 46% (n=17/37) (Supplemental Table S2). Mean HbA1c decreased from 8.3% ±1.0 to 8.0% ±1.1 (p-value = 0.044). Mean TAR >180 mg/dl decreased from 52.8±12.3 to 48.5±12.3 (p=0.011).
Table 1.
Change in glycemic and pump utilization outcomes 90 days pre/post U200-AIDa
| Pre-U200-AID | Post-U200- AID | p-value | |
|---|---|---|---|
| CGM Metrics (n=43) b | |||
|
| |||
| CGM active time (%) | 92.4 (75.5,97.1) | 93.6 (79.8, 97.7) | 0.267 |
| (days) | 83.2 (68.0, 87.4) | 84.2 (71.8, 87.9) | |
| Time in Range (70–180 mg/dL) | 44.6 (12.6) | 48.9 (11.4) | 0.012 |
| Time Above Range (TAR) (>180 mg/dL) | 54.2 (12.3) | 49.1 (13.2) | 0.003 |
| TAR (180 to 249 mg/dL) | 26.0 (23.8, 28.5) | 24.7 (22.4, 27.8) | 0.066 |
| TAR (>250 mg/dL) | 28.0 (11.9) | 24.2 (11.7) | 0.014 |
| Time Below Range (TBR) (<70 mg/dL) | 0.7 (0.2,1.4) | 0.8 (0.5, 1.7) | 0.053 |
| TBR (54 to 69 mg/dL) | 0.7 (0.2, 1.0) | 0.7 (0.4, 1.3) | 0.015 |
| TBR (<54 mg/dL) | 0.0 (0.0,0.3) | 0.0 (0.0,0.2) | 0.79 |
| Average Glucose (mg/dL) | 207.3 (27.2) | 194.1 (27.7) | <0.001 |
| Coefficient of Variation (%) | 37.7 (35.5, 39.4) | 39.2 (36.8, 42.7) | <0.001 |
|
| |||
| Metabolic Outcomes (n=34) c | |||
|
| |||
| Hemoglobin A1c (%) | 8.2 (1.1) | 8.0 (1.2) | 0.08 |
| Weight (kg) | 90.0 (78.1,106.8) | 98.8 (83.1,111.2) | <0.001 |
| BMI (kg/m2) | 32.3 (29.2, 34.4) | 33.4 (30.4,36.3) | <0.001 |
| BMI z-score | 2.1 (1.9,2.5) | 2.2 (2.0,2.5) | 0.006 |
|
| |||
| Pump metrics (n=37) d | |||
|
| |||
| Total daily insulin dose (units) | 102.6 (23.5) | 125.8 (38.9) | < 0.001 |
| Total daily insulin dose (units/kg/day) | 1.0 (0.9, 1.3) | 1.3 (1.0, 1.6) | < 0.001 |
| Daily basal insulin (units) | 52.7 (8.8) | 69.8 (19.6) | < 0.001 |
| Daily bolus insulin (units) | 48.5 (38.3,60.7) | 50.7 (40.0, 63.8) | 0.231 |
| Number of user-initiated boluses day | 4.6 (2.7, 5.5) | 3.5 (2.7, 5.3) | 0.248 |
| Time in automated mode (%) | 93.4 (78.2, 97.8) | 88.0 (73.3, 94.3) | 0.025 |
| Days between cartridge changes | 2.2 (0.5) | 3 (0.5) | < 0.001 |
Mean (SD), pre-post compared using Paired t-test. Otherwise, median (IQR) pre-post compared using Wilcoxon Signed Rank test
7 subjects excluded due to insufficient CGM data or discontinuing U200-AID within 14 days of initiation
9 subjects with missing data for A1c, Weight, BMI
Comparison only to those on U100-AID prior to U200-AID
When stratified by baseline HbA1c, those with HbA1c ≥8% (n=24) had a 7.3%-point increase in TIR (37.8%±8.8 vs 45.1% ±10.3, p<0.001), and 7.1%-point reduction in TAR >180mg/dL (61.0%±8.2 to 53.9±11.6, p=0.001) (Figure 1, Supplemental Table S3). For those with HbA1c <8% (n=16), Zone 1 TBR increased from 0.9% [IQR 0.2,1.6] to 1.2% [IQR 0.6,2.0] (p=0.016).
Figure 1. Waterfall plots showing change in CGM metrics for each participant in the period 90 days pre- and 90 days-post U200 initiation in AID with colors representing baseline HbA1c.
Figure 1A TIR: In stratified analysis, mean TIR (70–180 mg/dL) increased 37.8% (8.8) to 45.1% (10.3) overall (p < 0.001) in those with baseline HbA1c ≥ 8% and 53.5 (11.1) to 55.0 (11.7) (p = 0.603) in those with HbA1c < 8%. Figure 1B TAR: In stratified analysis, mean TAR (> 180 mg/dL) decreased from 34.5% (8.3) to 28.5% (10.8) (p = 0.006) in those with HbA1c ≥ 8%, and 18.9% (9.6) to 17.6% (11.2) (p = 0.511) in those with HbA1c < 8%.
Pump Utilization Metrics
Pump utilization metrics are reported only for those using U100-AID before initiating U200-AID (n=37, Table 1). Time in automated mode decreased (93.4% [IQR 78.2,97.8]) vs. 88.0% [IQR 73.3,94.3], p=0.025). The mean number of days between cartridge changes increased from 2.2±0.5 to 3.0±0.5 days (p<0.001) despite increases in TDD in total U100 units (102.6±23.5 vs. 125.8±38.9, p<0.001) and TDD by weight (1.0 [IQR 0.9,1.3] vs. 1.3 [IQR 1.0,1.6] units/kg/day, p<0.001). The daily basal insulin dose increased from 52.7± 8.8 to 69.8±19.6 units (p<0.001) but the number of user-initiated boluses and daily bolus insulin dose were unchanged (Table 1).
Safety
In the 365 days pre-U200-AID, there were no severe hypoglycemia events and three subjects experienced DKA (6 episodes/100 person-years). Over a median follow-up of 290 days (IQR 227,476) post-U200-AID, no subjects experienced DKA or severe hypoglycemia.
Discussion
AID has revolutionized treatment of T1D and is now standard of care (5, 6). However, the limited capacity of pump cartridges, which hold 1.6–3.0mL of insulin, equivalent to 160–300 units of U100 insulin, can act as a barrier to AID use for individuals with T1D and high insulin requirements (1). With the rising prevalence of obesity, expansion of AID use to type 2 diabetes, and trend towards smaller cartridge volumes to decrease pump size and on-body presence, strategies to improve AID accessibility and effectiveness for individuals with high insulin requirements are needed (7–10). In this real-world two-center retrospective cohort, AYAs safely used U200 in AID systems to manage T1D and experienced improvements in TIR and TAR without clinically meaningful changes in TBR. U200-AID was associated with a reduced frequency of cartridge changes despite an increased TDD. Over a median follow up period of 290 days, no DKA or severe hypoglycemia occurred while using U200-AID. Three (6%) AYAs discontinued U200-AID within 90 days of initiation for reasons unrelated to safety.
U200 has a comparable pharmacokinetic and pharmacodynamic profile to U100 and could facilitate use of AID in individuals with high insulin requirements (11, 12). However, the safety and effectiveness of U200-AID remain largely unexplored and U200 is not currently approved for use in AID. This is the first retrospective cohort study to evaluate the safety and effectiveness of U200-AID in T1D. Prior studies examining U200 in insulin pumps have been limited to case series or use in open-loop pumps. In a three-person case series, U200-AID decreased GMI without increasing hypoglycemia; however, this study only included one person with T1D and all subjects were adults(3). Other retrospective studies focused on use of concentrated U200 and U500 regular insulin (U500) in open-loop pumps and do not provide insights into safety and effectiveness in closed-loop algorithms (13, 14). Furthermore, U500 use is not ideal because it has a longer duration of action (18.3–21.5 hours) than U100 and U200 insulins which present theoretical safety concerns when used in AID (12, 15). In our cohort of overweight or obese AYA with median TDD >100 units/day, U200-AID resulted in a statically significant 7% increase in TIR in those with baseline HbA1c ≥8%; this change is clinically meaningful given that a 5% change in TIR is associated with a decrease in diabetes-related complications (16). Given the number of user-initiated boluses was unchanged, improved glycemia likely resulted from increased basal insulin delivery and longer duration between cartridge changes that may have prevented AYA from running out of insulin at unpredictable times of day. While improved bolusing habits would be the most ideal approach to improve glycemia, behavioral changes and medication adherence are frequently difficult in AYAs. Using U200 in AID where algorithms help attenuate hyperglycemia by increasing the basal rate or provide auto-boluses could be a more realistic approach to improving glycemia in this group. AYAs with baseline HbA1c <8% did not have statistically significantly improved TIR or clinically meaningful changes in TBR 54–69 mg/dL; however, the decreased cartridge change frequency has implications for quality of life and healthcare costs. Notably, an unexpected finding in this study was the increase in weight and BMI over a 90-day period. Due to the retrospective nature of our study, we cannot conclude how use of U200 may have changed weight trajectory. Regardless, these results underscore the importance of weight management in parallel with diabetes.
While the sample size was not powered to detect a difference in rare safety events, the absence of DKA and severe hypoglycemia is reassuring. Of note, our study could not capture other adverse events such as inadvertent use of U100 while on U200 profile settings. Inadvertent use of U100 could explain why some subjects had decreased TIR and percent time in automated mode, although this is unlikely given the increase in TDD. Furthermore, inadvertent use of U200 using U100 profiles could lead to hypoglycemia; however, in our study there was no clinically significant change in hypoglycemia. Given errors in calculating doses for U100 versus U200 has the potential to lead to two-fold differences in insulin dosing, patient education and careful documentation are crucial when using U200-AID. Additionally, providers must adjust delivery settings according to the proprietary algorithms for each AID system (see guidance in Supplemental Table S4).
Our study had several strengths including use of real-world data in AYA, a population that experiences higher rates of DKA, suboptimal technology use, and inadequate glycemic control(17). Limitations include a small sample size, short duration of follow up, and the use of limited AID systems. In our cohort, U200 was only used in CIQ and OP5; future studies on U200 use in other AID models, including Medtronic Minimed 780G and iLET bionic pancreas are needed. Additional limitations include lack of control group, potential U200/U100 misclassification, use of univariate analysis that prevented adjustment for confounding, and lack of data on anti-obesity medications or metformin. Potential confounding or effect modification were addressed by performing stratified analysis by baseline HbA1c and sensitivity analyses that only included those using U100-AID pre-U200-AID. In a sensitivity analysis that included only those previously using U100-AID, mean HbA1c improved by 0.3% and TIR increased by 3.0%, but the change in TIR did not reach statistical significance (46.3±12.2% vs. 49.3±10.7%, p =0.082). This could be related to the smaller sample size, as well as the possibility that the change from MDI to AID itself contributed to improvements in TIR in the primary analysis. Finally, complete-case analysis introduces the possibility of bias due to missing data.
Conclusions
In conclusion, in our cohort of AYA with T1D, U200-AID was associated with improved TIR and TAR without an increase in serious safety events. These findings highlight the potential of U200 to address barriers to AID use for individuals with high insulin requirements, particularly in those not attaining glycemic targets. Larger prospective randomized controlled trials using different types of AID are needed to further evaluate the safety and efficacy of U200 in AID.
Supplementary Material
Funding Statement
P.Y.C. is supported by NIH grant F32DK138739 and T32GM075766. BEM is supported by NIH grant K23DK129827. R.M.W is supported by NIH grant R01DK134955.
Author’s disclosures
B.E.M. has received industry sponsored research support from Medtronic, research supplies from Dexcom, Inc. and Digostics, and consulting and speaker’s fees from Insulet. R.M.W. is the site PI of sponsored clinical trials by Novo Nordisk, Lilly Diabetes and Sanofi, unrelated to this work.
Footnotes
Prior Presentation. Parts of this study were presented in abstract form at the Philadelphia Endocrine Society on May 22nd, 2024 and at the International Society of Pediatric and Adolescent Diabetes Annual Meeting Oct 16th to 19th 2024.
Conflict of Interest: No potential conflicts of interest relevant to this article were reported.
References
- 1.Sherr JL, Heinemann L, Fleming GA, Bergenstal RM, Bruttomesso D, Hanaire H, et al. Automated insulin delivery: benefits, challenges, and recommendations. A Consensus Report of the Joint Diabetes Technology Working Group of the European Association for the Study of Diabetes and the American Diabetes Association. Diabetologia 2023;66(1):3–22. [Google Scholar]
- 2.Linnebjerg H, LaBell ES, Dellva MA, Coutant DE, Leohr J. Bioequivalence of Ultra Rapid Lispro (URLi) U100 and U200 Formulations in Healthy Subjects. Diabetes Ther 2020;11(8):1709–20. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.AKCAN T, NEEDLEMAN L, BASINA M. 955-P: Effective Use of U-200 Insulin in OmniPod 5, Medtronic 780G, and Tandem X2 Systems. Diabetes. 2024;73(Supplement_1). [Google Scholar]
- 4.Battelino T, Danne T, Bergenstal RM, Amiel SA, Beck R, Biester T, et al. Clinical Targets for Continuous Glucose Monitoring Data Interpretation: Recommendations From the International Consensus on Time in Range. Diabetes Care. 2019;42(8):1593–603. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.American Diabetes Association Professional Practice C. 7. Diabetes Technology: Standards of Care in Diabetes-2025. Diabetes Care. 2025;48(Supplement_1):S146–S66. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Biester T, Berget C, Boughton C, Cudizio L, Ekhlaspour L, Hilliard ME, et al. ISPAD Clinical Practice Consensus Guidelines 2024: Diabetes Technologies: Insulin Delivery. Horm Res Paediatr 2024:1–34. [Google Scholar]
- 7.March CA, Becker DJ, Libman IM. Nutrition and Obesity in the Pathogenesis of Youth-Onset Type 1 Diabetes and Its Complications. Front Endocrinol (Lausanne). 2021;12:622901. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Kueh MTW, Chew NWS, Al-Ozairi E, le Roux CW. The emergence of obesity in type 1 diabetes. Int J Obes (Lond) 2024;48(3):289–301. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.PASQUEL FJ, DAVIS GM, HUFFMAN DM, PETERS AL, PARKER JC, LAFFEL LM, et al. 1904-LB: Glycemic Improvement with Use of the Omnipod 5 Automated Insulin Delivery System in Adults with Type 2 Diabetes—Results of the SECURE-T2D Pivotal Trial. Diabetes. 2024;73(Supplement_1). [Google Scholar]
- 10.Kudva YC, Raghinaru D, Lum JW, Graham TE, Liljenquist D, Spanakis EK, et al. A Randomized Trial of Automated Insulin Delivery in Type 2 Diabetes. N Engl J Med 2025. [Google Scholar]
- 11.de la Peña A, Seger M, Soon D, Scott AJ, Reddy SR, Dobbins MA, et al. Bioequivalence and comparative pharmacodynamics of insulin lispro 200 U/mL relative to insulin lispro (Humalog®) 100 U/mL. Clin Pharmacol Drug Dev 2016;5(1):69–75. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Umpierrez GE, Holt EH, Einhorn D, McGill JB. CONCENTRATED INSULINS: CLINICAL UPDATE OF THERAPEUTIC OPTIONS. Endocr Pract 2020;26(Suppl 3):1–12. [Google Scholar]
- 13.Lane WS, Weinrib SL, Rappaport JM, Przestrzelski T. A prospective trial of U500 insulin delivered by Omnipod in patients with type 2 diabetes mellitus and severe insulin resistance. Endocr Pract 2010;16(5):778–84. [DOI] [PubMed] [Google Scholar]
- 14.Kim G, Waldrep V. A Pediatric Case of Insulin Lispro U-200 as Continuous Insulin Infusion to Treat Type 1 Diabetes Mellitus Complicated by Insulin Resistance. AACE Clinical Case Reports. 2018;4(5):e419–e21. [Google Scholar]
- 15.de la Peña A, Riddle M, Morrow LA, Jiang HH, Linnebjerg H, Scott A, et al. Pharmacokinetics and pharmacodynamics of high-dose human regular U-500 insulin versus human regular U-100 insulin in healthy obese subjects. Diabetes Care. 2011;34(12):2496–501. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Battelino T, Alexander CM, Amiel SA, Arreaza-Rubin G, Beck RW, Bergenstal RM, et al. Continuous glucose monitoring and metrics for clinical trials: an international consensus statement. Lancet Diabetes Endocrinol 2023;11(1):42–57. [DOI] [PubMed] [Google Scholar]
- 17.Holman N, Woch E, Dayan C, Warner J, Robinson H, Young B, et al. National Trends in Hyperglycemia and Diabetic Ketoacidosis in Children, Adolescents, and Young Adults With Type 1 Diabetes: A Challenge Due to Age or Stage of Development, or Is New Thinking About Service Provision Needed? Diabetes Care. 2023;46(7):1404–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.