Longevity of the novel ConvaTec infusion set with Lantern technology

Abstract

Current insulin infusion sets are approved for only 2-3 days. The novel ConvaTec infusion set with Lantern technology is designed to extend infusion set wear time. The goal of this pilot study was to evaluate the duration of wear for this set. This was a pilot safety study in adults with type 1 diabetes using tethered insulin pumps. Participants inserted the set and wore it for 10 days or until failure. Among 24 participants, two were excluded. Forty-five per cent of the sets lasted 10 days. Median wear time was 9.1 (7.1, 10.0) days. Among 12 premature failures, six (50%) involved adhesive failures, four (33%) hyperglycaemia unresponsive to correction, one (8%) hyperglycaemia with ketones and one (8%) infection. Average CGM glucose per day of infusion set wear showed a statistically significant increase over time, while total daily insulin over the same period did not change. In this pilot study, the duration of wear for the novel infusion set exceeded previously reported commercial sets (P < .001). This extended wear technology may eventually allow for a combined glucose sensor and infusion set.

1 |. INTRODUCTION

The weak link in insulin pump and closed-loop therapy is the insulin infusion set that is inserted under the skin. There is limited information on the interaction between the infusion set, subcutaneous tissue and insulin. The infusion set can become occluded or kinked, and local inflammation and infection around the site can result in insufficient insulin delivery, high glucose levels, ketoacidosis and death.^1,2

There are few studies^3–7 published on the extended use of infusion sets beyond 3 days, although investigation is in progress by Capillary Biomedical (NCT04398030) and Medtronic (NCT04113694). Past studies have tested infusion sets for up to 7 days: Patel et al. reported a median wear time of 6.06 (4.35, 6.98) days,⁵ Karlin et al. a least squares mean wear time of 4.12-4.31 days,⁶ and Waldenmaier et al. a mean wear time of 6.2 ± 1.5 days.⁷ The study conducted by Patel et al. established failure criteria for infusion sets and was a randomized, open-label, crossover study of 20 subjects who wore Teflon catheter (Quick-Set) and steel catheter (Sure-T) infusions set for up to 1 week with no difference in the survival curves of the infusion sets. Length of wear of any infusion set was subject specific, and 30% of failures were secondary to uncorrectable hyperglycaemia, indicating partial catheter occlusion or limitations on insulin absorption in the subcutaneous space.⁵ The Karlin study was a multicentre, crossover trial that showed no difference in infusion set survival or mean glucose in 20 individuals with type 1 diabetes wearing the sets in lipohypertrophied and non-lipohypertrophied tissue.

Our hypothesis was that the ConvaTec infusion set with Lantern technology, a novel infusion set (Figure 1) with additional slitted openings designed to reduce the risk of occlusion and a modified coating, could increase wear time. The new hydrophilic coating mimics the cell membrane and is believed to reduce inflammation.

FIGURE 1.

ConvaTec infusion set with Lantern technology. A, Distal tip of novel ConvaTec infusion set with Lantern technology cannula (left), demonstration of increased surface area of distribution in unkinked state (middle) and catheter being kinked by underlying muscle with continued flow through side openings (right). B, Infusion set with no site failure at day 10 (top). Infusion set failure at day 8, with leakage (bottom)

2 |. METHODS

This pilot safety and extended wear tolerability study tested survival of the novel infusion set over 10 days of wear. The study was approved by the Stanford Human Subjects Research Compliance Office and was conducted in compliance with the standard of Good Clinical Practice and the Declaration of Helsinki. The study is registered on ClinicalTrials.gov (NCT03819634).

Adult participants (aged ≥22 years) were eligible if on tethered insulin pump therapy for at least 3 months using insulin lispro or aspart with an HbA1c of 9% or less. Participants needed to consume more than 60 g of carbohydrate each day (avoiding ketosis from diet) and be willing to follow the protocol. Participants were excluded if they were pregnant, had a hypoglycaemic seizure or severe diabetic ketoacidosis 3 months prior to screening, a cardiovascular event 6 months prior to screening, known tape allergies, active infection, active proliferative retinopathy, seizure disorder, inpatient psychiatric treatment, or the presence of other significant medical conditions such as adrenal disorders and renal failure. Written informed consent was obtained from all participants prior to performing any studyspecific procedures.

Potential participants were evaluated for eligibility through elicitation of a medical history, physical examination, point-of-care HbA1c and urine pregnancy tests in women. Eligible participants were then trained on study procedures, including use of the real-time continuous glucose monitor (CGM), blood glucose and ketone meter, and how to determine and document any infusion set failures. Participants were instructed to check blood ketones using a blood ketone meter if there was unexplained hyperglycaemia (high glucose levels outside of the usual times they may occur, such as in the initial few hours following a meal). Study staff trained participants on infusion set insertion and infusion sets were then placed by the participant. No additional adhesives could be used and there were no restrictions on activity. Participants wore a Dexcom G6 and their own insulin pump (either Tandem, Medtronic or Animas) and used insulin aspart or insulin lispro. Any automated delivery functions were disabled, such that participants were in open loop for the duration of the study. Infusion set failure was defined as follows (criteria 1-4 are considered occlusion events):

1. Ketones ≥0.6 mmol/L with glucose >250 mg/dL.

2. Unexplained hyperglycaemia, >250 mg/dL not responsive to corrections (approximately 50 mg/dL in 1 hour).

3. Occlusion alarm.

4. Infusion set kinking.

5. Fell off (adhesive failed).

6. Signs of infection (pain, redness, swelling).

Glucose events were determined by CGM and confirmed on a glucometer.

3 |. RESULTS

Twenty-four individuals agreed to enroll in the study at Stanford University from January to April 2019. During this study, no unresolvable pump occlusion alarms occurred. There was no ketoacidosis or severe hypoglycaemia events. Two participants were excluded. The first had a history of insulin leakage with all 90° 6-mm cannulas; she noted leakage with the experimental set and removed it. The second met failure criteria in less than 24 hours and did not replace the set as per study protocol. The remaining 22 participants were aged 40.1 ± 14.3 years with a diabetes duration of 26.4 ± 12.2 years; 12 (55%) identified as male and 10 (45%) as female, 18 (82%) identified as White with one (4.5%) identifying as Hispanic/Latino, and four (18%) as Asian; HbA1c was 7.1% ± 0.8%, total insulin 48.0 ± 16.7 units/day, and body mass index 28.0 ± 4.4 kg/m²; reported infusion set wear was 4 (3, 5) days; 19 (86%) used lispro while three (14%) used aspart; 10 (45%) used Medtronic (Dublin, Ireland) pumps, 10 (45%) used Tandem (San Diego, CA, USA) pumps and two (9%) used Animas (West Chester, PA, USA) pumps. The excluded subjects did not differ significantly from the included subjects.

Infusion sets were placed as follows: four (18%) left lower quadrant abdomen, two (9%) left upper quadrant abdomen, four (18%) right lower quadrant abdomen, four (18%) right upper quadrant abdomen, three (14%) left buttock, two (9%) right buttock, two (9%) left lower back and one (5%) left thigh. Insulin requirements during the trial varied from 20-125 units/day or 0.3-1.5 units/kg/day. Forty-five per cent of the novel infusion sets lasted the full 10 days (one with pus on removal). Median wear time was 9.1 (7.1, 10.0) days. Among the 12 premature set failures, six (50%) had adhesive failures on days 4.8/7.0/7.4/8.2/8.9/9.2, four (33%) had hyperglycaemia unresponsive to correction on days 5.2/6.5/6.8/7.8, one (8%) had hyperglycaemia with ketones of 0.8 mmol/L on day 6.4 and one (8%) had signs of infection on day 7.4. Figure 2A shows a survival curve for the novel infusion set. Eighty-two per cent of sets were still functional at 6.8 days. Figure 2B shows average CGM glucose per day of infusion set wear, showing a statistically significant increase over time. Of note, we include the partial final day of CGM data when some infusion sets failed because of hyperglycaemia. For infusion sets lasting 7 days, there was no significant increase in the daily mean glucose with extended wear. Total daily insulin is also plotted in Figure 2B, and did not change significantly.

FIGURE 2.

Performance of the novel ConvaTec infusion set. A, Survival curve by day for the novel ConvaTec infusion set with Lantern technology worn for a maximum of 10 days. Survival at 7 days is highlighted with a dotted line for comparison with prior studies, in which infusion sets were worn for a week. B, Average glucose on continuous glucose monitor (CGM) and total daily insulin (TDI) dosage by day of infusion set wear. At the bottom of each bar is the number of participants providing data. For CGM, participants had to provide sufficient CGM data and partial data for the last day of wear is included (when values may be higher with a failing infusion set). TDI was included on the final day only with more than 22 hours of insulin data. *P < .05 compared with mean glucose on day 1, all other comparisons in an ANOVA were insignificant

4 |. DISCUSSION

In this pilot feasibility study, the ConavaTec infusion set with Lantern technology exhibited significantly greater survival than commercial infusion sets. With survival capped at 7 days, median wear time (7.0 [7.0, 7.0] days) was significantly longer (P < .001) than in the 7-day trial conducted by Patel et al.⁵ Average CGM glucose increased over time, and this may have been driven by data on the day of failure. The total daily insulin data were excluded if less than 22 hours of insulin data were available, as may occur on the day of failure. Therefore, if there was an infusion set failure associated with hyperglycaemia, we may detect the increase in average glucose but not total daily insulin. The rate of occlusion events at 7 days was 18%, lower than the 30%-35% published in prior studies.^5,6 This was not statistically significant, and was possibly attributable to the limited sample. Larger studies are planned to establish whether the novel design decreases occlusion events.

The experimental device presents an opportunity for a 7-day wear combination CGM and infusion set. With the advent of automated insulin delivery (AID), patients are required to wear both a sensor and insulin infusion set. Wearing two devices is a potential barrier to the use of these systems. Out of 1006 people wearing a CGM, 27% discontinued CGM use after 1 year and 33% explained that they did not want to have devices on two sites of their body.⁸ Finding multiple, comfortable sites can be particularly challenging with younger children, who have a smaller body surface area. When the sensor is in one location and the infusion set is in another, it means two separate insertions for the patient, and frequent reapplication of adhesive tapes to keep both devices in place. When the adhesive tape is removed a layer of epidermis is also removed, which is one of the natural barriers to infections. Current infusion sets are approved for 2-3 days of wear,^4,9 while sensors are approved for 7-14 days of wear with greater accuracy after the first day.¹⁰ Thus, any attempt to create a combined sensor and insulin infusion set requires an infusion set with greater longevity. Several groups have attempted to combine a sensor and infusion set into one device, but it was not practical because of limited infusion set longevity.^11–14 Recent work by Tschaikner et al. reveals that off-the-shelf CGM and infusion sets can be combined by running the sensor wire through the cannula.¹³ With the sensor wire extending 6 mm through the infusion set cannula opening, the sensor glucose obtained from the single-site device did not differ significantly from a separately worn CGM.¹⁴

The strengths of the current study include real-world use of a new infusion set. The participants encompassed users of different insulins, insulin pumps with different body compositions and experiences with infusion sets. Insulin glulisine and Fiasp were excluded at the time of this study, as they did not have approval in all pumps. The pilot observational study design had limitations, it was not statistically powered and there was no control infusion set. This is particularly important as prior studies have revealed that the individual is the best predictor of subsequent infusion set failures.⁵ The reported usual wear time for this group was 4.1 days, which is slightly longer than the recommendation of 2-3 days. It could be the case that participants in this cohort could wear sets for longer than those in prior studies. Additionally, infusion set failure criteria were based upon previous infusion set studies and regulatory guidance. In particular, the criteria may miss some partial occlusion and the specific signs of infection used may not catch every occurrence. Only the 6 mm cannula length was available for this study, which limited use in the participant with leakage from shorter cannulas. Future studies should also include longer cannula length. Because of regulatory constraints, we were unable to test the performance of the infusion sets with AID and active predictive low glucose suspend, because of concerns that the intermittent suspension of insulin delivery would increase the risk of catheter occlusion. AID has now become the standard for commercial insulin pumps. We plan to perform a 7-day masked, randomized controlled crossover study comparing the novel ConvaTec infusion set with comparable, commercially available infusion sets, without slits or coating in systems using AID.

DATA AVAILABILITY STATEMENT

The data that support the findings of this study are available from the corresponding author upon reasonable request.