Development of the Extended Infusion Set and Its Mechanism of Action

Abstract

Continuous subcutaneous insulin infusion (CSII, or insulin pump) and continuous glucose monitoring (CGM) sensors have been increasingly used and associated with improved glycemic control by people with type 1 diabetes and insulin-requiring type 2 diabetes. Commonly used infusion sets in most CSII systems are limited to a wear time of 3 days. In contradistinction, CGM sensors are currently approved for seven and more days of wear. With the motivation to provide a 7-day infusion set that matches the CGM wear time and to improve patient experience, the recently CE-marked and FDA 510k–cleared Medtronic extended infusion set (EIS) was designed.s The EIS offers enhanced new features that include use for up to 7 days, improved convenience, comfort, and better quality of life for insulin pump users.

Since continuous subcutaneous insulin infusion (CSII) with an external pump was introduced in persons with type 1 diabetes (T1D) in the 1970s, there have been progressive increases in the number of pump users. When used properly, the treatment is considered to be safe and effective.^1-3 Major challenges of CSII include infusion set (IS) failure ⁴ which can lead to hyperglycemia and diabetic ketoacidosis^5,6 and local tissue reactions including scarring and infection. ⁴ The recommendation for changing ISs every 2 to 3 days, published in 1983, was based on limited case reports related to infections. ⁷ Since then, there have been several published trials assessing IS wear duration which support limiting the use of aforementioned ISs currently on the market to 2 to 3 days.^8-12

With the advent of real-time CGM (RT-CGM), many groups have worked on using CSII and RT-CGM to achieve partial or full closed-loop (CL) control of insulin delivery. Current systems require that patients wear an IS at one site and a RT-CGM sensor at a second site. Wearing devices at two sites for different durations is seen as a deterrent to wide-scale adoption of the CL system. Finding suitable sites on which to attach the devices can be challenging, particularly for children. Combining the IS with the glucose sensor on a single platform would reduce the number of devices worn and site insertions. As most CGM sensors are approved for 7 to 14 days of wear, extending the IS wear duration was a significant unmet need. This technology report describes the development of the recently CE-marked and FDA 510k–cleared (K210544) ¹³ Medtronic extended infusion set (EIS), approved for up to 7-day wear (Figure 1).

Figure 1.

Illustration of pump-reservoir-extended infusion set (EIS) system (left) and the EIS components (right).

IS Failure Modes During Longer Wear Durations

To develop the EIS, failure modes associated with extending the wear duration of 2/3-day sets in published studies were examined.^8-12,14 Data from past studies were somewhat consistent: neither the type of insulin (Novolog vs Humalog) nor the IS type (steel vs Teflon) was a significant factor in determining IS wear time. No IS extended-wear study was performed using Apidra as insulin glulisine is only approved for use in a pump system for 2 days. ¹⁵ The study conducted by the Stanford University group ¹⁰ had the most detailed analyses of failure modes (Table 1), which was benchmarked for EIS development in 2015. In this study, the most dominant 7-day failure mode was the unexplained hyperglycemia (UH) at 30% failure rate and the strongest predictor of prolonged 7-day IS function was the individual subject. Unexplained hyperglycemia being the dominant failure mode was confirmed by the later studies.^16-18 The failure modes associated with the prolonged IS use were summarized by Dr Buckingham in DTM 2021 Symposium (Table 1). ¹⁹

Table 1.

Seven-Day Failure Modes for 2- to 3-Day Infusion Sets in the Benchmark Study ¹⁰ and for EIS in the Pivotal Design Validation Study.

	2/3-Day IS 10 (Benchmark)	% in literature 19 (Mean, range)		7-Day EIS 13
7-Day failure rate a	64%	52% (34%-67%)		25%
Infusion set failure modes b	Observed		Likely or Identified Causes	Observed
Unexplained hyperglycemia	30%	22% (6%-30%)	Site occlusion resulted from biological responses to insertion injury, cannula, insulin, or other foreign materials Gel/aggregates inside the cannula, mechanical cannula kinking	0.7%
Discomfort (pain, erythema, edema) at infusion site	23%	10% (3%-14%)	Biological responses to insertion injury, cannula, adhesive, insulin, or other foreign materials	1.6%
Accidental pull-outs	10%	7% (1%-13%)	Accidents	7.3%
Adhesive failure	5%	8.5% (5%-14%)	Adhesion loss due to wear and tear	6.2%
Infection	4%	2.5% (0%-5%)	Biological response to microbial ingression, loss of preservative	0.5%
Others	Not reported	6.5% (5%-8%)	Miscalculating the time, subject personal decision of removal, and reasons unrelated to the IS	11.8%

Abbreviations: EIS, extended infusion set; IS, infusion set.

^aThe rate of IS failure is defined as the number of IS removals associated with the failure mode divided by total number of sets inserted.

^bFailure modes might be overlapped, as subjects could indicate multiple reasons for one removal.

Although UH has been reported in many CSII clinical studies,^8-12,16 the underlying causes of UH remain poorly understood. Early studies disclosed by Novo Nordisk ²⁰ hinted that insulin stability might be an important contributor to the UH. Later, the in vitro chemical testing results by DeFelippis et al ²¹ and Sharrow et al ²² demonstrated 7- to 14-day insulin stability in a pump system under aggressive conditions of temperature and agitation, which presumed that UH was not due to insulin instability during pumping. Some hypotheses have theorized that preservatives might be the root cause for loss of infusion site viability over time.^23-26 However, other studies^27-29 have supported the initial finding by Novo Nordisk that the main contributor to UH is not the preservatives but insulin stability in the pump/reservoir/IS system. Other chemicals (eg, silicone oil) have also been shown to aggravate insulin aggregation in the fluid path. ³⁰

In recent years, in vitro/in vivo studies conducted by us confirmed that, although temperature and agitation during pumping had no significant impact on insulin chemical stability, they caused formation of insulin aggregates/fibrils, which were highly inflammatory as tested by macrophage cell culture. ³¹ Preservative is also very important for insulin formulation stability. Preclinical study using insulin formulated with lower preservative content (≤50% Label Claim) induced more insulin aggregates in pumped infusates, significantly reduced IS wear duration and resulted in more infusion site/cannula occlusion in the T1D porcine model. ³¹ These findings permitted us to focus on maintaining the insulin stability in the fluid path to extend IS wear duration. ³²

EIS Design and Mechanism of Action

Figure 1 illustrates the principal EIS components and its use in a Medtronic insulin pump system. As presented in Table 2, the main EIS design focus was to maintain insulin stability in the fluid path. Other failure modes listed in Table 1 were also addressed. The key EIS features include a proprietary polyvinyl alcohol (PVA) foam and filter with proper pore size in the fluid path. Polyvinyl alcohol is known to stabilize insulin conformationally and thermodynamically.^33,34 Our internal study data confirmed the literature findings that the released insulin, after interacting with PVA foam, were more stable in the fluid path. In addition, the foam/filter combination effectively removes particulates/aggregates (≥5 µm) and hydrophobic molecules (eg, lubricant from the reservoir) and reduces site inflammation caused by these foreign molecules. ³² Furthermore, the EIS tubing is optimized for preventing preservative loss, therefore better maintaining insulin formulation stability. ³¹ Engineering and clinical user studies were conducted to select an adhesive patch for improving skin adhesion and wear comfort as well as well as keeping dermal layer injury to minimum, especially during an accidental pull. ³⁵ Finally, a predicate state-of-art insertion device with soft Teflon cannula at 90° insertion angle was incorporated to ensure insertion reliability, reduce injury/inflammation, and increase comfort. ³⁶

Table 2.

Principal Components in the EIS and Their Main Functions.

Key components in EIS	Design rationale	Benefit provided	Targeted failure mode
In-line PVA foam	PVA is known to stabilize insulin conformationally and thermodynamically.33,34 In addition, the foam can serve as a coarse filter for removing larger particulates and hydrophobic molecules.	Prevents aggregation and thermodynamically stabilizes insulin.	Unexplained hyperglycemia Discomfort Infection
In-line filter	Particulates (≥10 µm) and hydrophobic chemicals in insulin infusate can affect insulin stability and induce inflammatory tissue responses. The EIS in-line filter is selected to remove particulates ≥5 µm and hydrophobic molecules.	Removes particulates, aggregates, and hydrophobic substances.
Multilayer tubing	A trilayer medical tubing was thoughtfully designed to reduce gas permeability and mitigate preservative loss. Also, the tubing has a lubricious inner layer for insulin flow.	Mitigates preservative loss in the fluid path and reduces insulin aggregation by reducing gas permeability
New adhesive patch	The adhesive patch was selected based on the material features: nonsensitizing, easy to handle and convert, breathable and conformable, sterilizable, and provides sufficient adhesion over extended wear by engineering and user clinical studies.	Provides breathability and comfortability, whereas maintaining extended wear time	Adhesive failure Discomfort
6-mm or 9-mm soft Teflon cannula at 90° insertion angle	Soft cannula can reduce tissue injury during extended wear. 37 Selection of cannula length and insertion angle was based on the amount of subcutaneous tissue presented in the target users. A 90° angle is chosen to shorten the needle and therefore shorten the injury path in tissue. Both 6- and 9-mm cannula lengths are provided for users with thinner and thicker subcutaneous tissue layers.	Minimizes tissue injury, reduces cannula material in the subcutaneous tissue, reduces the fat lobules impacted, ensures insulin delivery reliability	Unexplained hyperglycemia Discomfort
State-of-art insertion device	The preloaded insertion device with automatic needle retraction was leveraged from Mio Advance infusion set (510k K173879). The insertion needle and soft cannula are hidden from the user before, during and after insertion of the soft cannula.	Ensures consistent insertion depth, reduces insertion failure, increases comfort, reduces injury and inflammation

Abbreviations: EIS, extended infusion set; PVA, polyvinyl alcohol.

Design Validation

The new device was investigated in a pivotal design validation study that enrolled 259 participants (48% men, 45.0 ± 14.1 years), including 132 Humalog users and 127 Novolog users. Detailed data for the EIS pivotal trial have been published in Diabetes Technol Ther. ³⁸ and FDA K210544 summary. ¹³

This single-arm, multicenter, nonrandomized trial was conducted in individuals with T1D using their own MiniMed 670G systems. Each participant was given 12 EIS devices, with each worn for ≥174 hours or until set failure. The intention-to-treat (ITT) population had a total of 3041 EIS wears. This study demonstrated that the EIS, when worn for up to 7 days, was safe and rated with high satisfaction, without adversely affecting glycemic control in adults with T1D. ³⁸

Including all reasons, EIS 7-day failure rate was 25.2% with the failure modes examined and also presented in Table 1. The top three failure modes were Others, 11.8%; Accidental Pull-outs, 7.3%; and Adhesive Failure, 6.2%. In the previous studies, the “Other” category might have been counted as “Scheduled Removal” or not counted as those were not device-related failures. The EIS failure rates for the other two modes were similar to the corresponding means observed in literature. Improving adhesive longevity and comfortability is needed to further reduce EIS 7-day failure rate.

The rate of EIS failure due to UH (the main target of EIS design) at the end of day 7 (168 hours) was <1%, as presented in Table 3 for those using Humalog or Novolog in the ITT population. Results from the EIS pivotal study validated the design intents as described in Table 2. There was no predictor of prolonged 7-day IS function and no significant difference was observed between Humalog and Novolog users in any of the clinical outcomes evaluated.

Table 3.

Rate of Infusion Set Failure Due to Unexplained Hyperglycemia at the End of the Day 6 and Day 7.

Wear time (at the end of)	Insulin used in ITT population	Number of subjects	Number of infusion sets	Number of failed infusion sets due to unexplained hyperglycemia (number of subjects)	Rate of infusion set failure a	Prespecified threshold (upper CL)	Actual result (upper CL)	Success criteria met?
Day 6 (Primary Endpoint)	Humalog	132	1561	1 (1)	0.06%	20%	0.45%	Yes
	Novolog	127	1480	4 (4)	0.27%	20%	0.71%	Yes
Day 7 (Secondary Endpoint)	Humalog	132	1561	2 (2)	0.13%	20%	0.51%	Yes
	Novolog	127	1480	6 (5)	0.41%	20%	1.00%	Yes

Abbreviations: CL, closed loop; EIS, extended infusion set; IS, infusion set; ITT, intention to treat; UH, unexplained hyperglycemia.

^aThe rate of IS failure is defined as the number of EIS removals associated with UH divided by total number of sets inserted.

Excluding “Not for Cause” failures, the overall survival rate of the EIS for 7 days was 77.8%, comparable with the survival rate of 3-day sets used for 3 days in the benchmark study. ¹⁰

Conclusion

In conclusion, recent availability of the EIS, which comprises a thoughtfully designed fluid path that allows insulin infusion with less particulates/aggregates and hydrophobic chemicals, may reduce site inflammation and site pain for CSII users. This, in addition to use for up to 7 days, improved convenience and comfort, can help provide better quality of life for insulin pump users.