Even though it is largely a matter of definition, the term “insulin pump therapy” must be expanded to include a series of innovations.
Beginning in 2007, systems for continuous glucose monitoring (CGM) were connected to insulin pumps with the goal of gradually automating “classical” insulin pump therapy. As a result, various forms of insulin pump therapy have emerged:
SuP represents an intermediate step between classical CSII and CSII with AID, forming a hardware connection between an insulin pump and a CGM system.
The main purpose of SuP is to prevent hypoglycemia by interrupting insulin delivery when glucose levels reach a preset hypoglycemia threshold (typically 70 mg/dL or 3.9 mmol/L) or predictively, when CGM trends indicate that glucose levels are approaching this threshold.
In the version where insulin delivery is suspended after reaching the hypoglycemia threshold, an alarm is triggered first.
If the patient does not respond, insulin delivery is automatically stopped for a set period (e.g., 120 minutes). After this period, the pump automatically resumes insulin delivery unless the patient has already restarted it manually (LGS – Low Glucose Suspend).
Nevertheless, mild hypoglycemia may still occur—for example, if the threshold is set at or below the actual hypoglycemia level.
In the predictive suspension variant, the advantage is that hypoglycemia has not yet occurred at the time of interruption, but it can now be prevented.
The suspension and subsequent resumption of insulin delivery both occur automatically once the risk has passed (see Fig. T1D8).
If alarm notifications are disabled, the patient may not even notice these automatic adjustments.
Fig. T1D8: Effect of Predictive Low Glucose Suspension in a Person with Type 1 Diabetes
The suspension of insulin delivery occurred in a timely manner whenever glucose levels were falling (as seen in the glucose profile above; the resulting red-dashed areas below indicate the periods of insulin suspension). This led to a stable glucose pattern at a good glycemic level, without episodes of hypoglycemia.
The PubMed database contains more than 470 English-language articles reporting clinical results achieved with SuP.
As an example of the success of hypoglycemia suspension, one pediatric study in children with an average age of 11.5 years can be cited [1].
Algorithms also include safety modes that operate during automated control.
These ensure user safety and prevent the system from exiting automatic mode too easily when metabolic conditions are uncertain or glucose data are missing.
Specific implementations vary across systems, but the underlying safety principles are similar.
Further details can be found in the “AID system profiles” (AID-Steckbriefe) published by the Working Group on Diabetes Technology (AGDT):
“Steckbriefe für Systeme zur automatisierten Insulin-Dosierung (AID)” – AGDT.
It is recommended that these profiles be available in diabetes care practices.
Fig. T1D9: Functioning of Different Algorithms
The first studies using hybrid AID systems already demonstrated that people with type 1 diabetes achieved HbA1c levels around 7% or lower, with the proportion of glucose values within the target range of 70–180 mg/dL (3.9–10 mmol/L) — the so-called Time in Range (TIR) — averaging above 70% [4].
Importantly, these improvements were achieved without episodes of severe hypoglycemia.
Another essential finding is that similar results have been observed not only in clinical trials of various AID systems but also in large real-world cohorts of people with type 1 diabetes.
For example, an analysis of 119,274 patients in the United States and 4,690 in Europe showed an average TIR of about 72% and an HbA1c just below 7% [5].
With the use of Advanced Hybrid Automated (AH-AID) systems, the outcomes are even slightly better, owing to the automatic administration of correction boluses.
Evidence indicates that glycemic outcomes achieved with AID systems surpass those obtained with other therapeutic options in large patient populations.
As a result, acute complications (such as severe hypoglycemia and diabetic ketoacidosis) as well as long-term diabetes complications become significantly less frequent.
This, of course, applies primarily to people with type 1 diabetes who were able to adopt such technology early in their disease course.
A particularly important aspect is that these favorable outcomes are achieved without excessive therapeutic effort from the users.
The body of evidence supporting the successful use of insulin pump therapy with AID is exceptionally strong, with approximately 1,100 published studies [6].
As an example, a meta-analysis [7] examined 28 randomized controlled trials involving all known AID systems.
Results were compared with other forms of subcutaneous insulin therapy using CGM.
Across all relevant studies, people with type 1 diabetes using AID systems showed:
Fig. T1D10: Results from a Meta-Analysis of 28 Randomized Controlled Trials
AID systems improved mean glucose levels by more than 25 mg/dL and increased Time in Range (TIR) by over 21% [27].
Once the decision to begin insulin pump therapy has been made, the next question is which insulin pump should be used.
In most cases, the decision will clearly favor an AID system. It has become relatively uncommon for people with diabetes to choose insulin pump therapy without AID functionality.
The first decision criterion is therefore:
In principle, the performance of various AID systems in terms of metabolic control is similar — all have demonstrated positive results in clinical studies.
Therefore, when considering the complete AID system (i.e., insulin pump plus CGM), the future user should first choose the system that best suits them personally, since they will be wearing it continuously.
Besides the CGM system, several external factors of the insulin pump should be considered:
External and Design Factors
Regulatory and Eligibility Criteria
Technical Aspects to Consider
For example, the Omnipod patch pump must be replaced every three days and holds a maximum of 200 units of insulin — this must align with the user’s daily requirement.
AID systems can be operated not only in automatic mode but also manually, with or without CGM connectivity.
Thus, manual settings remain important, especially when using the pump in classical CSII mode.
Programming and Control Options
Final Recommendation
It is always advisable to discuss the choice of AID system or insulin pump with the diabetes care team.
These professionals usually have hands-on experience with multiple systems — an invaluable resource when making an informed decision.
[1] Biester T, Kordonouri O, Holder M, et.al. “Let the Algorithm Do the Work”: Reduction of Hypoglycemia Using Sensor-Augmented Pump Therapy with Predictive Insulin Suspension (SmartGuard) in Pediatric Type 1 Diabetes Patients. Diabetes Technol Ther. 2017 Mar;19(3):173-182.
[2] Thomas A, Heinemann L. Algorithms for Automated Insulin Delivery: An Overview. J Diabetes Sci Technol. 2022;16(5):1228-1238.
[3] Thomas A. AID: So wird die Insulinabgabe berechnet. Diabetes-Forum 5/2021: 10-15.
[4] Bergenstal RM, Garg S, Weinzimer SA, et.al.. Safety of a hybrid closed-loop insulin delivery system in patients with type 1 diabetes. JAMA 2016;316:1407-1408.
[5] Agrawal P, Stone M, Gopalakrishnan S et.al. Real-World Data from the MiniMed™ 670G System Commercial Launch. ADA 2018, Orlando, Diabetes 2018; 7(Suppl.1): A249-A250.
[6] Datenbank Pubmed, Stichwort “automated insulin delivery”. Letzter Aufruf: 03.06.2025.
[7] Di Molfetta S, Di Gioia L, Caruso I, et.al. Efficacy and Safety of Different Hybrid Closed Loop Systems for Automated Insulin Delivery in People With Type 1 Diabetes: A Systematic Review and Network Meta-Analysis. Diabetes Metab Res Rev. 2024; 40(6):e3842.