Abstract
Introduction
Insulin pump manufacturers are making automated insulin dosing ubiquitous, and this is likely to increase the uptake of insulin pump technology. The weak link in continuous subcutaneous insulin infusion (CSII) is still the infusion set or cannula, which serves as the interface between the pump and the body. Many of the highlighted articles this year touch on this crucial piece of technology. Buckingham and colleagues published a pilot study regarding the novel ConvaTec infusion set with Lantern technology, as well as the pivotal trial for Medtronic Extended-Wear Infusion Set (also manufactured by ConvaTec), which has now been marked by The Conformitè Europëenne (CE) and approved by the United States Food and Drug Administration (FDA). As highlighted in the 2020 yearbook, Tschaikner and colleagues have demonstrated that an all-in-one infusion set and continuous glucose monitor (CGM) single-port device is technically feasible. The next steps are human trials of these combination devices. For those using more than 43 units of U-100 insulin per day, synchronizing changes of the insulin reservoir with the 7-day infusion sets will either require a larger capacity or, more likely, pump compatibility with concentrated insulin. The team at the University of Graz elucidated changes in pressure and resistance that occur over 7 days' wear. They demonstrate that the insulin protein itself is the primary driver of these changes. They suggest uses for pressure sensing that have previously been described by Diatech Diabetes (Memphis, TN). Finally, Kanapka and colleagues from the Jaeb Center for Health Research analyzed data regarding hyperglycemic infusion set failure in the Control-IQ pivotal trials. Recapitulating data from an animal study presented by Eisler in the 2020 yearbook, the authors report less hyperglycemic failure among those using an angled Teflon catheter. Ninety-degree Teflon infusion sets were used the most in the trial, and it may be prudent to recommend angled Teflon sets in clinical practice.
Another form of insulin delivery hardware is the implanted intraperitoneal insulin pump. He and colleagues present an outstanding analysis of the root cause of catheter failure. As evidence mounts for the deleterious effect of insulin protein on subcutaneous delivery, it was unknown if intraperitoneal catheter occlusion was driven by insulin or foreign body response. The findings suggests that while insulin cannot be excluded, the primary driver is foreign body response. Thus, novel catheter technology must be employed to expand access to this promising technology. In the world of smart insulin pen technology, Ekberg and colleagues used a NovoPen 6 to document the number of missed basal doses and its effect on glycemia. The group found that even one missed basal dose per week can have a deleterious impact.
The remaining literature on the use of insulin pumps among those living with type 1 diabetes highlighted the need for equity in access to insulin pump therapy. Fantasia and colleagues revealed significant disparities in diabetes technology use at their Boston safety-net hospital. White patients with type 1 diabetes had more access to insulin pumps and CGM even after adjusting for socioeconomic factors. In Canada, Song and colleagues report that insulin pump access is more common in regions with government funded programs and among the wealthier. Neves reports on predictors of effectiveness of insulin pump among their Portuguese type 1 diabetes cohort.
With regard to type 2 diabetes, Freckmann and colleagues provided a review of studies published between 2003 and 2018 on the use of insulin pumps. As with type 2 diabetes itself, there was significant heterogeneity in the studies, making them very difficult to compare. With the tremendous evolution of noninsulin agents, the role of insulin pumps is evolving, and it becomes increasingly important to elucidate the insulin secretory capacity of study participants.
Gagouri and colleagues retrospectively evaluated factors associated with pump discontinuation among those with both type 1 and type 2 diabetes in their French clinic. They emphasize that these should not be used to restrict dissemination of this valuable technology, but rather to suggest those who need more help. Concluding on the issue of safety, Biester and colleagues report that patch pump therapy was associated with lower percentage of patients with DKA per year and lower percentage of patients with a level 3 hypoglycaemia per year compared to multiple daily injections (MDI) after 3 years in the Diabetes Patienten Verlaufsdokumentation (DPV) registry. This real-world data also showed a much higher baseline ketoacidosis rate in the pump group, which declined over time.
Key Articles Reviewed
Ekberg NR, Hartvig NV, Kaas A, Møller JB, Adolfsson P
Lal RA, Hsu L, Zhang J, Schøndorff PK, Heschel M, Buckingham B
He J, Renard E, Lord P, Cohen D, Cem-Duranty E, Place J, Gu B, Wang X, Yenduri G, Burgess DJ
Regittnig W, Tschaikner M, Tuca AC, Simic A, Feiel J, Schaller-Ammann R, Licht AH, Jungklaus M, Pieber TR
Kanapka LG, Lum JW, Beck RW
Brazg R, Garg SK, Bhargava A, Thrasher JR, Latif K, Bode BW, Bailey TS, Horowitz BS, Cavale A, Kudva YC, Kaiserman KB, Grunberger G, Reed JC, Chattaraj S, Zhang G, Shin J, Chen V, Lee SW, Cordero TL, Rhinehart AS, Vigersky RA, Buckingham BA
Fantasia KL, Wirunsawanya K, Lee C, Rizo I
Song C, Booth GL, Perkins BA, Weisman A
Neves JC, Neves JS, Neves C, Carvalho D
Freckmann G, Buck S, Waldenmaier D, Kulzer B, Schnell O, Gelchsheimer U, Ziegler R, Heinemann L
Biester T, Schwandt A, Heidtmann B, Rami-Merhar B, Haak T, Festa A, Kostow S, Müller A, Mönkemöller K, Danne T, DPV Initiative
Gargouri I, Hadja Inna AA, Franc S, Picaud P, Penfornis A, Amadou C
SMART INSULIN PENS
Smart Pen Exposes Missed Basal Insulin Injections and Reveals the Impact on Glycemic Control in Adults with Type 1 Diabetes
Ekberg NR1,2, Hartvig NV3, Kaas A3, Møller JB3, Adolfsson P4,5
1Center for Diabetes, Academic Specialist Center, Stockholm, Sweden; 2Department of Molecular Medicine and Surgery, Karolinska Institutet, Karolinska University Hospital, Stockholm, Sweden; 3Novo Nordisk A/S, Bagsværd, Denmark; 4Department of Pediatrics, The Hospital of Halland, Kungsbacka, Sweden; 5Institute of Clinical Sciences, Sahlgrenska Academy at University of Gothenburg, Gothenburg, Sweden
Background
As of 2020, only 28% of adults in Sweden living with type 1 diabetes and receiving subspecialty diabetes care achieved HbA1c goals. Of these individuals, 74% were on an injection regimen. With the advent of smart insulin pens, it is now possible to track basal insulin administration in this population and document its effect on glycemic control.
Methods
This was a post hoc analysis of a one-arm prospective real-world study conducted at 12 diabetes clinics across Sweden from May 2017 to April 2019 and included adults with type 1 diabetes using an insulin degludec smart pen (NovoPen 6). The sample was restricted to those with ≥70% CGM data and two basal injections over a 14-day period. Missed basal insulin was defined as ≥40 hours between two basal injections. Estimates of probabilities of missed basal insulin doses were based on a cumulative linked mixed model, with participants as random effects. Associations between missed basal doses and glycemic outcomes were explored. As a covariate, the number of missed bolus insulin doses was determined for participants by using a connected bolus insulin pen. A missed bolus was stipulated to occur −15 to +60 minutes relative to a meal detected by the glucose rate increase detector (GRID) method (1).
Results
Thirty-two participants provided 315 fourteen-day periods for analysis. The estimated 14-day probability of no missed basal doses was 78%; one missed basal dose, 17.7%; two missed basal doses, 3.6%; and more than two missed basal doses, 0.6%. Missed basal insulin doses were significantly associated with higher mean glucose, higher glucose management indicator values, and lower time in range. Similar results were observed when adjusted for missed boluses.
Conclusions
These data reveal the challenges that exist with administering insulin degludec, even with access to smart insulin pens and CGM. One missed basal dose per week can result in a significant change in glycemic control.
Comments
Even with the relatively long dosing window afforded by insulin degludec, 22% of those in the study missed at least one dose in a 14-day window. This is particularly significant in a closely monitored clinical trial population with access to ultra-long–acting insulin, smart pens, and CGM. Those with less access to these products are likely to have even more missed basal doses. Even more significant, missed basal insulin is a risk factor for absolute insulinopenia and resulting diabetes-related ketoacidosis. This risk of ketoacidosis is mentioned in the article, but the observed rate in this cohort is not reported. There is a need for automated reminders and systems for detecting missed doses. While there are a multitude of imperfect strategies for the detection of missed/late-meal boluses from CGM and other available data (2 –13), detecting missed basal doses is relativity straightforward to implement.
INSULIN INFUSION
Longevity of the Novel Convatec Infusion Set with Lantern Technology
Lal RA1,2,3, Hsu L1, Zhang J2, Schøndorff PK4, Heschel M4, Buckingham B1,3
1Division of Endocrinology, Department of Pediatrics, Stanford University School of Medicine, Stanford, CA; 2Division of Endocrinology, Department of Medicine, Stanford University School of Medicine, Stanford, CA; 3Stanford Diabetes Research Center, Stanford, CA; 4ConvaTec, Copenhagen, Denmark
Background
The objective of this study was to assess the wear time of the novel ConvaTec infusion set with Lantern technology (ConvaTec, Copenhagen, Denmark) with a modified coating and additional slitted openings designed to reduce the risk of occlusion and inflammation.
Methods
This was a pilot study enrolling adult participants with an HbA1c of 9% or less on tethered insulin pump therapy for at least 3 months using insulin lispro or aspart. Participants were instructed to check blood ketones using a blood ketone meter if there was unexplained hyperglycemia. 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 continuous glucose monitor (CGM) and their own insulin pump (either Tandem, Medtronic, or Animas) with automated delivery functions disabled. Infusion set failure was defined as follows: ketones ≥0.6 mmol/L with glucose >250 mg/dL, unexplained hyperglycemia >250 mg/dL not responsive to corrections (approximately 50 mg/dL in 1 hour), occlusion alarm, infusion set kinking, infusion set falling off (adhesive failed), and signs of infection (pain, redness, and swelling).
Results
A total of 22 participants completed the study and had the following characteristics: age, 40.1±14.3 years; HbA1c, 7.1±0.8%; diabetes duration, 26.4±12.2 years; body mass index, 28.0±4.4; and 12 (55%) were male. Insulin requirements varied from 20 to 125 units/day or 0.3 to 1.5 units/kg/day. Of the novel infusion sets, 45% lasted the full 10 days (one with pus on removal). The median wear time was 9.1 (IQR, 7.1–10.0) days; 82% of sets were still functional at 6.8 days. Among the 12 premature set failures, six (50%) had adhesive failures on days 4.8, 7.0, 7.4, 8.2, 8.9, and 9.2; four (33%) had hyperglycemia unresponsive to correction on days 5.2, 6.5, 6.8, and 7.8; one (8%) had hyperglycemia with ketones of 0.8 mmol/L on day 6.4; and one (8%) had signs of infection on day 7.4. 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.
Conclusions
In this pilot study, the duration of wear for the novel infusion set exceeded previously reported commercial infusion sets, with 82% of sets still functional at 6.8 days.
Comments
The insulin infusion set remains the weak link in insulin pump and closed-loop therapy. 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 (often without any alarms), ketoacidosis, and death (14). Current insulin infusion sets are approved for only 2 to 3 days, but many users anecdotally report that they do not change infusion sets at recommended intervals. Consequently, there is a need for more reliable insulin infusion sets with extended wear time. Previous studies have demonstrated wear times between 4.1 to 6.2 days (15 –17). In this pilot study, the authors demonstrate a median wear time of 9.1 (IQR, 7.1–10.0) days, with 82% of sets still functional at 6.8 days. When the CGM data from the final day of the cannula use were included, the authors found a statistically significant increase in mean CGM glucose over days 8, 9, and 10. However, for infusion sets lasting 7 days, there was no significant increase in the daily mean glucose with extended wear. Authors speculate that this experimental device presents an opportunity for a 7-day wear combination of CGM and infusion set. Limitations of the current study include not testing ultra-fast-acting insulin such as Fiasp and Lyumjev (not approved at the time of the study), the use of only 6-mm cannula, and a small sample size. Failure criteria may miss some partial occlusion and some instances of infection and inflammation. Consequently, there is a need for further studies, including the full spectrum of rapid- and ultra-rapid-acting insulins and different cannula lengths.
Root Cause Determination of Intraperitoneal Catheter Obstructions: Insulin Amyloid Aggregates vs Foreign Body Reaction
He J1, Renard E2,3, Lord P4, Cohen D4, Cem-Duranty E2, Place J3, Gu B1, Wang X1, Yenduri G1, Burgess DJ1
1Department of Pharmaceutical Sciences, University of Connecticut, Storrs, CT; 2Department of Endocrinology, Diabetes and Nutrition, Montpellier University Hospital, Lapeyronie Hospital, Montpellier, France; 3Institute of Functional Genomics, CNRS, INSERM, University of Montpellier, Montpellier, France; 4PhysioLogic Devices, Inc., Alpine, CA
Background
Continuous intraperitoneal insulin infusion is limited by catheter obstruction of unknown etiology. Current silicone catheters consist of a polyethylene inner tubing (inner diameter 0.2 mm), silicone outer tubing (outer diameter 2.79 mm), and the silicone tip, where the insulin is released. Clinically, these catheter failures are described as follows: tip blockage from fibrin and insulin tip encapsulation (colloquially “sock”) from fibrotic tissue, giant cells, and insulin catheter encapsulation from fibrotic tissue surrounding the peritoneal portion of the catheter lumen occlusion from insulin aggregates
There are 6.2 years of data reflecting 544 patient-years of intraperitoneal therapy. The data revealed that 52.4% of individuals had no catheter obstruction, 33.7% of individuals had one catheter obstruction, 11.5% of individuals had two catheter failures, and 1.9% of individuals had three catheter obstructions. Among 65 catheter obstructions, 52 (80%) were deemed secondary to tip blockage and 13 (20%) to catheter encapsulation.
Methods
A 2-year clinical investigation was conducted along with the development of a novel rodent model to examine the mechanisms behind intraperitoneal catheter obstruction. The clinical protocol included collection of insulin samples from the pump reservoirs and explanted catheters of patients at Montpellier University who were using implanted intraperitoneal pumps. Human samples were sectioned and stained with H&E, trichrome, toluidine blue (mast cell visualization), and Congo red (for insulin amyloids). Sprague Dawley rats with and without streptozotocin-induced diabetes were implanted with a catheter designed to mimic the Medtronic MMT4024 intraperitoneal catheter but with a proximal injection port and shorter peritoneal catheter component (17.8±0.7 cm in human vs 4.5±0.3 cm for rats). Commercial syringe pumps were connected to the injection ports and used to deliver either Sanofi U400 Insuman Insulin (5 units/kg) or commercial diluent (containing the same excipients but no insulin) twice per week at constant pressure. Explanted catheters were sectioned and stained with H&E, trichrome, and Congo red. Immunofluorescence was conducted for macrophages and fibroblasts.
Results
Twenty-eight explanted catheters (20 from occlusion and 8 from system change) were collected between June 6, 2017, and April 19, 2019. Laparoscopic diagnosis revealed tip blockage, tip encapsulation, and catheter encapsulation. Twelve explanted catheters were evaluated histologically: 2 demonstrated tip encapsulation while 10 showed tip blockage that was either partial (Type I) and complete (Type II). All catheters examined revealed macrophages, spindle-shaped fibroblast-like cells, collagen, and few neutrophils. The change in composition from Type I to Type II to tip encapsulation followed the typical pattern of foreign body reaction. Tissue irritation appeared to be a contributor to fibrotic tissue buildup. Congo red staining did reveal amyloid aggregates in most but not all samples, indicating the obstruction likely preceded the insulin aggregation. In the rodent model, tip blockages of the implanted catheters were present after only 3 days, and most of the rats had fully blocked catheters after 2 weeks; in 8 to 12 weeks, the tip blockages grew into the encapsulation tissue. There were no insulin amyloid aggregates in the tip blockages, and a similar progression of obstruction occurred regardless of the use of insulin or diluent. Tip blockage tissue was found in the peritoneal cavity of rats, indicating that, in many cases, the obstructions were dislodged. Compared to healthy rats, those with streptozotocin-induced diabetes had a reduced host response to implanted catheters. Rats with severe peritoneal irritation exhibited greater collagen accumulation.
Conclusions
Taken together, these findings suggest insulin aggregates are unlikely to be the leading cause of catheter obstruction and that a foreign body tissue response is primarily responsible for intraperitoneal catheter failure. This foreign body reaction may be worsened by perturbation of the adjacent tissues or organs. One cannot rule out the possibility that insulin aggregates in the infusion system could accelerate this foreign body reaction. Possible mitigations include the use of catheters with enhanced biocompatibility, configurations that reduce foreign body reactions (such as smooth round tip) or anti-inflammatory drug eluting catheters.
Comments
As highlighted in the ATTD 2021 yearbook, Swinney and colleagues implicated the insulin protein in unexplained hyperglycemia resulting from extended wear of subcutaneous insulin infusion sets in a porcine model of diabetes (18). Whereas the subcutaneous failures occur over a week, intraperitoneal catheters fail over years with a reported incidence of 11.9 failures per 100 patient-years. One might hypothesize that intraperitoneal catheter obstruction is a result of the insulin protein, foreign body reaction, or both. Although implanted intraperitoneal insulin pumps have been used for decades, the authors are the first to provide compelling evidence for the underlying mechanism of catheter obstruction. It is important to note that the current standard of care for implanted intraperitoneal pump users requires insulin refills every 6 to 12 weeks and time-consuming rinses with sodium hydroxide every 6 months. For wider dissemination of this technology, rinses will have to be eliminated. Because the participants in the study received standard care, it is unclear whether eliminating the rinses would precipitate a greater number of insulin-induced failures. Regardless, catheter technology has advanced significantly since the MMT4024 was first developed, and future intraperitoneal catheters can benefit significantly from the knowledge gained as a result of this work.
Insulin Induces a Progressive Increase in the Resistance of Subcutaneous Tissue to Fluid Flow: Implications for Insulin Pump Therapy
Regittnig W1, Tschaikner M1, Tuca AC1, Simic A1, Feiel J2, Schaller-Ammann R2, Licht AH3, Jungklaus M1, Pieber TR1,2
1Department of Internal Medicine, Division of Endocrinology and Diabetology, Medical University of Graz, Graz, Austria; 2Joanneum Research Forschungsgesellschaft mbH, HEALTH-Institute for Biomedicine and Health Sciences, Graz, Austria; 3Roche Diabetes Care GmbH, Mannheim, Germany
Background
The magnitude of subcutaneous infusion set-related flow resistance largely depends on the diameter and length of the infusion set's components, the viscosity of insulin, and the frictional forces that the subcutaneous tissue exerts. As the fat cells are virtually impenetrable to the flow of fluids, insulin flow is largely confined to the narrow pores of the extracellular matrix (∼25 nm). Prior studies have offered conflicting data regarding tissue-related flow resistance with saline. The aim of the present study was to determine the effect of insulin on tissue-related flow resistance among people with type 1 diabetes.
Methods
Participants with type 1 diabetes on insulin pump therapy and with an HbA1c <86 mmol/mol (10%) were asked to wear two Accu-Chek Spirit Combo insulin pumps with Accu-Chek FlexLink infusion sets inserted on the abdomen or hips for 7 days. Participants were allowed to select cannula and tubing length. One pump was filled with insulin aspart and the other with insulin diluting medium. Participants were asked to match infusions of insulin and diluting medium. Daily measures of pressure required to infuse boluses of diluting medium were performed with a DPT-100 disposable pressure sensor.
Results
Thirty participants aged 18 to 64 years with type 1 diabetes were included in the study. On day 1, maximum pressure and tissue flow resistance were similar for both insulin and diluting medium (P>.20). For diluting medium, during the subsequent study days, these measures remained similar to those on day 1 (P>.13). For insulin, pressure and resistance progressively increased with duration of wear. By the end of day 7, maximum pressure reached a geometric mean of 25.8 kPa (geometric SD 2.11), and tissue flow resistance a geometric mean of 8.64 kPa·s/μL (geometric SD 3.48). These represent a 3.5- and a 20.6-fold increase relative to the values observed on day 1 (P<.001).
Conclusions
These results suggest an insulin-protein–induced progressive increase in subcutaneous tissue flow resistance.
Comments
The present study adds to a growing body of evidence that the insulin protein itself impedes delivery to the subcutaneous space. The authors postulate that (a) insulin fibrils and/or precipitates may deposit in the pores of the extracellular matrix, progressively increasing tissue-related flow resistance or (b) that collagen fibrils and cells accumulate at the tissue adjacent to the infusion site, diminishing the area for fluid flow. Uniquely, the authors found that all 30 participants wore the sets for the full 7 days with no infection or hyperglycemic failure despite a 3.5-fold increase in pressure and 20.6-fold increase in tissue flow resistance. There were failures on insertion that were mitigated by choosing a different cannula length. Additionally, the authors report adhesive failure that was treated with additional adhesive. Without this intervention, these may have eventually been classified as hyperglycemic failure. Insulin leakage is likely underappreciated and more clearly understood, given the progressive loss of adhesion and significant increases in pressure and resistance that occur over time.
The authors report several important use cases for pressure monitoring that have also been proposed by Diatech Diabetes (Memphis, TN;
Insulin Pump Infusion Set Failures Associated with Prolonged Hyperglycemia: Frequency and Relationship to Age and Type of Infusion Set During 22,741 Infusion Set Wears
Kanapka LG1, Lum JW2, Beck RW1
1Jaeb Center for Health Research, Tampa, FL; 2Artificial Pancreas Coordinating Center, Jaeb Center for Health Research, Tampa, FL
Background
Failures of subcutaneous insulin pump infusion sets are commonly reported. Several studies have assessed the failure rate of sets beyond the labeled wear time. An analysis of a large dataset under controlled conditions and according to labeling is needed.
Methods
Data were analyzed from two Control-IQ clinical trials, in which 263 participants aged 6 to 72 years used 22,741 infusion sets. The authors define hyperglycemic failure as CGM measuring >300 mg/dL immediately before removal and >250 mg/dL continuously for at least 2 hours before removal with at least 90 minutes >300 mg/dL out of the prior 120 minutes. Differences in failure rates among age groups and infusion set types were evaluated.
Results
Among 22,741 infusion sets, 748 (3.3%) were removed before 72 hours in association with prolonged hyperglycemia. The percentage replaced for prolonged hyperglycemia within 48 hours was 1.8% and 1.0% within 24 hours. Mean duration of continuous time >250 mg/dL before removal was 5.1±3.7 h. Using a less restrictive definition of failure related to hyperglycemia, 1688 (7.4%) sets were removed before 72 hours with a glucose level >300 mg/dL at the time of removal. The frequency of insulin set failure with prolonged hyperglycemia was lower in adults ≥18 years old (1.9%) than in those 14 to 17 years old (5.8%, P<.001) or 6 to 13 years old (4.4%, P=.002). The 90-degree Teflon sets, which were the most frequently used, had the highest frequency of prolonged hyperglycemia failure within 72 hours (4.0%) when compared with the angled Teflon set frequency (1.3%, P=.01) and the steel set frequency (1.9%, P=.006).
Conclusions
Based on the data from these 22,741 infusion sets, infusion-set changes associated with prolonged hyperglycemia occur on average about 4 times a year, with the frequency being higher in youth than in adults. The frequency also appears to be higher with straight Teflon sets than with angled Teflon sets or steel sets.
Comments
This post hoc analysis of the Control-IQ pivotal trials provides valuable information on the rate of hyperglycemic infusion set failure. Importantly, all users were using an automated insulin delivery system, which would reduce hyperglycemia if the set were functioning. The authors, therefore, developed a classification of hyperglycemic failure dependent on CGM and that differed from those previously used in extended-wear trials (15 –17,19,20). As this was a pivotal study, all participants were compelled to wear infusion sets as labeled for 2 to 3 days, and, indeed, mean wear duration was 2.9±1.3 days. Cumulative failure probability increased as a function of time, and it is likely real-world wear time may exceed the wear time reported in this trial. Prior work has highlighted the potential benefit of angled Teflon infusion sets in a porcine model of diabetes (21). Given this supportive clinical data, it may be prudent to recommend using an angled Teflon set for those on pump therapy.
Evaluation of Extended Infusion Set Performance in Adults with Type 1 Diabetes: Infusion Set Survival Rate and Glycemic Outcomes from a Pivotal Trial
Brazg R1, Garg SK2, Bhargava A3, Thrasher JR4, Latif K5, Bode BW6, Bailey TS7, Horowitz BS8, Cavale A9, Kudva YC10, Kaiserman KB11, Grunberger G12, Reed JC13, Chattaraj S14, Zhang G14, Shin J14, Chen V14, Lee SW14, Cordero TL14, Rhinehart AS14, Vigersky RA14, Buckingham BA15
1Rainier Clinical Research Center, Renton, WA; 2Barbara Davis Center for Diabetes, Aurora, CO; 3Iowa Diabetes and Endocrinology Center, West Des Moines, IA; 4Medical Investigations, Inc., Little Rock, AR; 5AM Diabetes and Endocrinology Center, Bartlett, TN; 6Atlanta Diabetes Associates, Atlanta, GA; 7AMCR Institute, Escondido, CA; 8Metabolic Research Institute, Inc., West Palm Beach, FL; 9Diabetes and Endocrinology Consultants of Pennsylvania, Feasterville-Trevose, PA; 10Mayo Clinic, Rochester, MN; 11SoCal Diabetes, Torrance, CA; 12Grunberger Diabetes Institute, Bloomfield Hills, MI; 13Endocrine Research Solutions, Roswell, GA; 14Medtronic, Northridge, CA; 15Stanford University School of Medicine, Palo Alto, CA
Background
Standard insulin infusion sets are labeled for replacement every 2 to 3 days to avoid set failure. Since 2017, Medtronic has been designing and testing an extended-wear infusion set manufactured by ConvaTec. The novel infusion set includes a new adhesive patch, H-cap connector, and tubing with a fluid path design that improves insulin preservative retention and stability. These changes were based on observations that insulin degradation and preservative loss increase inflammatory response. This pivotal trial evaluated the safety and performance of a new extended-wear infusion set when used for 7 days by adults with type 1 diabetes.
Methods
This single-arm, pivotal trial evaluated the safety and performance of the new extended-wear infusion set when used for 7 days with 12 consecutive wears by adults with type 1 diabetes using the MiniMed 670G system in Auto Mode with insulin lispro or insulin aspart. Safety endpoints included incidence of serious adverse events, serious adverse device effects, unanticipated adverse device effects, severe hypoglycemia, severe hyperglycemia, diabetes-related ketoacidosis, and skin infection. The failure rate due to unexplained hyperglycemia, the overall survival rate, HbA1c, mean sensor glucose time spent in established glucose ranges, total daily insulin delivered, and satisfaction with the infusion set were determined.
Results
The intention to treat population (N=259, 48% men, 45.0±14.1 years) wore a total of 3041 novel infusion sets. Overall rates of serious adverse events, severe hypoglycemia, severe hyperglycemia, and infection were 3.8, 2.5, 104.1, and 20.1 events per 100 participant-years. The rate of extended-wear infusion set (EIS) failure due to unexplained hyperglycemia at the end of day 7 was 0.1% (95% CI, 0.03–0.51) for insulin lispro and 0.4% (95% CI, 0.16–1.00) for insulin aspart. At the end of day 7, the overall survival rate for the infusion set was 77.8% (95% CI, 76.2–79.3), glycemic control did not change, and participants reported greater satisfaction with the extended-wear infusion set compared with standard insulin infusion sets worn before the study (P<.001).
Conclusions
The extended-wear infusion set, when worn for up to 7 days, was deemed safe by the authors and rated with high satisfaction by participants without adversely affecting glycemia in adults living with type 1 diabetes.
Comments
The data provided in this single-arm study was sufficient to grant FDA approval for a 7-day wear time and sets a precedent for the approval of future extended-wear infusion sets. Critical choices by the sponsor and agency include the lack of control arm, use of automated insulin dosing system, and, unlike in past trials, basing failure around unexplained hyperglycemia alone. Prior studies suggest the individual wearing the infusion set is the best predictor of future infusion set failure (15). Therefore, there is always some concern for selection bias in infusion set survival studies. Other infusion set failure events, such as occlusion alarms, cannula kinking, adhesive failure, infection, and leakage, were not included in the primary outcome. The definition of infusion set failure and management has varied between sites conducting the study, such that it becomes very difficult to compare results between trials. In becoming a precedent for extended-wear infusion set regulatory approval, one would hope that uniform criteria may be established. Future studies are likely to use calibration-free CGM, and it would be useful to define failure based on CGM metrics to avoid finger sticks. With the approval of an extended-wear infusion set, work has already commenced on a single port CGM and infusion set. For individuals using more than 43 units of U-100 insulin per day, wearing a set for 7 days means changing the reservoir asynchronously from the infusion set. Larger reservoirs or compatibility with concentrated insulin analogs (e.g., U-200) will be needed to synchronize infusion set and reservoir changes.
INSULIN PUMPS IN TYPE 1 DIABETES
Racial Disparities in Diabetes Technology Use and Outcomes in Type 1 Diabetes in a Safety-Net Hospital
Fantasia KL, Wirunsawanya K, Lee C, Rizo I
Section of Endocrinology, Diabetes, and Nutrition, Boston University School of Medicine and Boston Medical Center, Boston, MA
Background
Limited data exist regarding diabetes technology use among adults with type 1 diabetes in urban racially/ethnically diverse safety-net hospitals. Safety-net hospitals are institutions that deliver health care and other health-related services to patients who are uninsured or underinsured, including those covered by Medicaid, regardless of a patient's ability to pay. In this setting, the authors examined racial/ethnic differences in CGM and pump use.
Methods
A retrospective review of 227 patients ≥18 years of age with type 1 diabetes seen in an urban safety-net endocrinology clinic during 2016 and 2017 was completed (mean age, 39; 80% English-speaking; 50% had public insurance). Diabetes technology use (defined as use of CGM, pump, or both) and clinical outcomes were examined by race/ethnicity.
Results
A total of 227 patients with type 1 diabetes were included in the study: 97 White (43%), 57 Black (25%), 35 Hispanic (15%), and 38 other (17%). Overall, the mean age was 38.8±13.0 years, 59% were male, the mean duration of diabetes was 20.9±12.8 years, and the most recent mean HbA1c was 8.9% ±2.0% (74±22 mmol/mol). Fifty percent of White patients had an annual household income of more than $75,000; only 19% of Black patients, 17% of Hispanic patients, and 27% of patients who identified as other were in that income category.
Overall, 30% of patients used CGM and 26% used pumps. After adjusting for age, language, insurance, and annual income, diabetes technology use in non-White patients were significantly lower than in White patients; it was predominantly lower in Black patients (adjusted OR 0.25 [95% CI, 0.11–0.56]) and patients who identified as other race/ethnicity (adjusted OR 0.30 [95% CI, 0.11–0.77]). At the highest household income level (≥$75,000/y), Black and Hispanic individuals were significantly less likely than White individuals to use diabetes technology (P<.0007).
Among all patients, White patients had the lowest mean HbA1c: 8.1%±1.5% (65±16 mmol/mol) in White patients, 10.1%±2.2% (87±24 mmol/mol) in Black patients, 9.2%±2% (77±22 mmol/mol) in Hispanic patients, and 8.9%±1.9% (74±21 mmol/mol) in patients who identified as other race/ethnicity (P<.0001). Glycemic control was better in patients using any form of diabetes technology when compared with those not using technology (8.1%±1.6% [65±18 mmol/mol] vs 9.4%±2.1% [79±23 mmol/mol], P<.0001). CGM and pump users had the lowest HbA1c across all racial/ethnic groups. Additionally, CGM-only users were found to have lower HbA1c than pump-only users, except among Black individuals; among Black patients who used any form of diabetes technology, CGM-only users had the highest HbA1c.
Conclusions
In an urban safety-net setting, more than half of White patients with type 1 diabetes used diabetes technology, whereas less than one-third of non-White patients used diabetes technology. Racial/ethnic disparities in diabetes technology use and glycemic control were observed even after adjusting for sociodemographic factors. Further research should explore barriers to accessing diabetes technology in non-White populations.
Comments
Previous studies have shown that racial and ethnic minority youth with type 1 diabetes have lower rates of diabetes technology use than their White counterparts (22,23). However, access to diabetes technology and health outcomes for adults of racial and ethnic minority groups with type 1 diabetes remains understudied. This study examines the use of diabetes technology in a cohort of adults attending an urban racially/ethnically diverse safety-net hospital clinic in the United States. Notably, despite the availability of adequate health insurance and even after adjusting for multiple confounders, technology use was significantly lower in non-White patients. In contrast to prior research, sociodemographic factors such as annual household income and language spoken do not appear to be the primary drivers of these disparities. So, what are the possible reasons for this? While this study cannot answer this question, the authors speculate on several possible underlying mechanisms.
Insurers in the United States and other funding bodies and local commissioners outside the United States often have essential requirements regarding self-management behaviors to qualify for funding diabetes technology devices. These self-management behaviors often require significant resources, time, and effort of individuals. These requirements may systematically disadvantage lower socioeconomic status and racial and ethnic minority patients, whose economic and social conditions may restrict their actions at home or work. Thus, removing restrictive eligibility criteria may enable increased access to diabetes technology, which can serve as a tool to self-motivate, engage in self-care, and improve health outcomes. Authors also highlighted other contributory factors including conscious and unconscious bias, structural racism, and unmeasured factors.
Limitations of this study include having data from only one center and a relatively small sample size for this kind of observational study. Nevertheless, its findings highlight the need for further research to understand underlying factors and patient, provider, and organizational barriers that contribute to disparities in the application of diabetes technologies across socioeconomically diverse groups. Sustained efforts are required to ensure a more equitable allocation of diabetes technology and to minimize health-care disparities.
Impact of Government-Funded Insulin Pump Programs on Insulin Pump Use in Canada: A Cross-Sectional Study Using the National Diabetes Repository
Song C1, Booth GL1,2,3,4, Perkins BA1,2,5, Weisman A1,4,5
1Department of Medicine, University of Toronto, Toronto, Ontario, Canada; 2Institute of Health Policy, Management, and Evaluation, University of Toronto, Toronto, Ontario, Canada; 3Li Ka Shing Knowledge Institute of St Michael's Hospital, Toronto, Ontario, Canada; 4ICES, Toronto, Ontario, Canada; 5Lunenfeld-Tanenbaum Research Institute, Mount Sinai Hospital, Toronto, Ontario, Canada
Background
Insulin pump access in type 1 diabetes may be inequitable. The authors studied the association between government funding programs for insulin pumps and rates of insulin pump use and studied the disparities between pump users and nonusers in five Canadian provinces.
Methods
Adults with type 1 diabetes aged 18 to 55 were identified in the National Diabetes Repository, a primary care electronic medical record database of individuals with diabetes from five Canadian provinces covering about 78% of the Canadian population. Government funding was available in Ontario, Alberta, and Newfoundland (if age ≤25 in Newfoundland) and absent for Manitoba, Quebec, and Newfoundland (if age >25 in Newfoundland). Proportions of individuals using insulin pumps were compared between provinces with and without pump funding programs. Multivariable logistic regression models were used to estimate the odds of insulin pump use, adjusting for confounders. Univariate logistic regression models were used to estimate the odds of insulin pump use per predictor and pump funding program status.
Results
Of the 95,699 individuals with an electronic medical record encounter between January 1, 2015, and June 30, 2019, in the National Diabetes Repository, 1559 were identified as adults with type 1 diabetes. Of these, 1320 lived in provinces with government-funded insulin pump programs (631 pump users), and 239 lived in provinces without government-funded insulin pump programs (90 pump users). Individuals living in provinces with and without insulin pump funding programs were of similar age (40 years, IQR: 31–49 vs 41 years, IQR: 33–39; P=.33) and sex distribution (47% male vs 49% female, P=.53). Of 1559 adults with type 1 diabetes, proportions using insulin pumps were 47.8% (95% CI, 45.1%–50.5%) and 37.7% (95% CI, 31.5%–44.1%) in provinces with and without pump funding programs (P=.0038). Adjusting for age, sex, HbA1c, income quintile, and rural/urban location, the odds ratio for insulin pump use was 1.45 (1.08–1.94) for provinces with pump funding programs compared with provinces without. Higher income was associated with a greater odds of insulin pump use in provinces with pump funding programs, and rural/urban location was not associated with insulin pump use.
Conclusions
Insulin pump use is more common in regions with government funding programs, and having a higher income in those provinces with funding programs was also associated with insulin pump use. Further research is required to understand better and comprehensively address persistent income disparities between pump users and nonusers despite the availability of reimbursement programs.
Comments
As already highlighted, despite proven benefits and cost-effectiveness, inequitable access to insulin pump therapy remains. In this cross-sectional study, individuals living in provinces with government-funded insulin pump programs were approximately 50% more likely to use an insulin pump. While authors hypothesized that government-funded insulin pump programs might remove financial barriers to insulin pump use, they found a significant residual association of higher income with insulin pump use even in provinces with pump funding. Strengths of this study include it being a natural experiment within the same country with differing provincial policies for insulin pump funding. Limitations include the inclusion of only individuals aged 18 to 55 years and possible overestimation of pump users (46% of the type 1 diabetes population). Given the benefits of insulin pump therapy and the emerging promise of closed-loop systems (which require both insulin pumps and continuous glucose monitoring devices) (24), further strategies for improving the equitable use of insulin pumps are needed.
Predictors of the Effectiveness of Insulin Pumps in Patients with Type 1 Diabetes Mellitus
Neves JC1,2, Neves JS1,2,3,4, Neves C1,2,4, Carvalho D1,2,4
1Department of Endocrinology, Diabetes and Metabolism, Centro Hospitalar Universitário de São João, Porto, Portugal; 2Faculty of Medicine, University of Porto, Porto, Portugal; 3Department of Surgery and Physiology, Cardiovascular Research Unit, Faculty of Medicine, University of Porto, Porto, Portugal; 4Institute for Research and Innovation in Health Sciences, University of Porto, Porto, Portugal
Background
Insulin pump therapy has become the preferential treatment for type 1 diabetes because it mimics the physiological secretion of insulin better than multiple daily injections. However, not all patients improve with insulin pump therapy. This study aims to determine the predictors of the effectiveness of insulin pumps in type 1 diabetes.
Methods
This was a single-center retrospective observational study of patients with type 1 diabetes who started insulin pumps between 2005 and 2020, were followed for at least 6 months after starting insulin pumps, and were 18 years old or older in the last time point assessed. All patients had access to an automatic bolus calculator, and no patient was using sensor-augmented pumps during the evaluation. Hypoglycemia was defined as the occurrence of episodes of glucose <70 mg/dL reported in the clinical record, and severe hypoglycemia was defined as hypoglycemia requiring external assistance for recovery. Data from four time points (before starting pump and at 6, 12, and 36 months after starting pump) were evaluated for glycemic control and safety outcomes. The association of baseline predictors with outcomes was analyzed by using linear and logistic regression models.
Results
The authors evaluated 136 patients (57.4% women, age 36±12 years, duration of type 1 diabetes 14±9 years). During the follow-up, there was a mean decrease of HbA1c of 0.9±1.2%. The improvement in HbA1c was independent of sex, age, and duration of type 1 diabetes. Higher baseline HbA1c, family history of diabetes, and not being treated with statins were predictors of improvement in HbA1c. Eighty percent had improved HbA1c, 75.4% had improved HbA1c without worsening hypoglycemia, 23.5% had improved hypoglycemia, and 18.4% had improved hypoglycemia without the worsening HbA1c. Not being treated with statins and having a higher baseline HbA1c predicted improved HbA1c without worsening hypoglycemia. History of hypoglycemia was a predictor of severe hypoglycemia. Family history, higher baseline HbA1c, and psychological/psychiatric disorders were predictors of ketoacidosis.
Conclusions
The benefits of insulin pumps were independent of sex, age, and duration of type 1 diabetes. Baseline HbA1c, family history of diabetes, treatment with statins, history of hypoglycemia, and psychological/psychiatric disorders were predictors of outcomes. These findings may allow the identification of patients who benefit most from insulin pump therapy or are at increased risk of complications.
Comments
In keeping with previous studies and meta-analysis, (25,26) authors found that those with higher baseline HbA1c had a more significant drop in HbA1c. In addition, the improvements in HbA1c were independent of sex, age, and duration of type 1 diabetes. The current study's sample size is relatively small (N=136) and single center in origin. The authors assessed a relatively large number of baseline associations with key outcomes of interest. Another limitation is the lack of sensor use. Since virtually everybody living with type 1 diabetes has episodes of mild hypoglycemia (with and without pumps), it is not entirely clear how robust the assessments of hypoglycemia outcomes are, especially if this was assessed as a binary outcome. Perhaps one unexpected finding is the association of statin therapy with lower HbA1c improvements after starting pump therapy. Since statins are a very important class of medication in the primary and secondary prevention of cardiovascular disease, this observation merits further evaluation in appropriately designed studies in multiple daily injection and insulin pump therapies. It is important to note that observational studies do not prove causality, and yet unmeasured confounders are possible. For example, as the authors correctly discuss, statin use could be a marker for another mechanism, such as a dietary or lifestyle factor. In this study, 4.5% of the population had ketoacidosis during the 3 years of follow-up. Higher baseline HbA1c levels and the presence of psychological/psychiatric disorders were predictors of having diabetes-related ketoacidosis after insulin pump therapy, once again highlighting the need for extra education to minimize the occurrence of this life-threatening complication after starting insulin pump therapy.
INSULIN PUMPS IN TYPE 2 DIABETES
Insulin Pump Therapy for Patients with Type 2 Diabetes Mellitus: Evidence, Current Barriers, and New Technologies
Freckmann G1, Buck S1, Waldenmaier D1, Kulzer B2, Schnell O3, Gelchsheimer U4, Ziegler R5, Heinemann L6
1Institut für Diabetes-Technologie, Forschungs- und Entwicklungsgesellschaft mbH an der Universität Ulm, Ulm, Germany; 2Forschungsinstitut Diabetes Akademie Bad Mergentheim, Germany; 3Forschergruppe Diabetes e.V., Helmholtz Zentrum, Munich, Germany; 4Roche Diabetes Care GmbH, Mannheim, Germany; 5Diabetes Clinic for Children and Adolescents, Münster, Germany; 6Science Consulting in Diabetes GmbH, Neuss, Germany
Background
Insulin pump therapy is well established in type 1 diabetes. In contrast, the role of insulin pump therapy in type 2 diabetes is evolving. Despite the advent of novel treatments such as SGLT-2 inhibitors and incretin-based therapies, many people with type 2 diabetes do not reach modern glycemic targets and may need addition of insulin therapy, particularly those with lower insulin reserves as opposed to those with insulin resistance and high insulin levels. Insulin therapy in type 2 diabetes can be basal or rapid-acting insulin only, twice-daily premixed insulin, basal-bolus therapy, or insulin pump therapy.
Methods
This article is a narrative review. The authors summarize results from studies published between 2003 and 2018, including uncontrolled, retrospective, and randomized controlled trials. In addition, the authors discuss the merits of simpler pumps for type 2 diabetes. Barriers and solutions, as well as future research areas, are also discussed.
Results
The studies in the article differ considerably regarding their designs, methods, insulin pumps, and study populations, making it difficult to compare results directly. The trials had different objectives: changing from multiple daily injection (MDI) to pump, short-term pump use in patients newly diagnosed with type 2 diabetes, pump in hospitalized settings, and overnight-only pump. The groups that were studied included newly diagnosed, poorly controlled, pump- or insulin-naive patients with type 2 diabetes, and patients with type 2 diabetes in general. Additionally, the number of participants recruited for these studies ranged from 10 to 609 subjects, and the pump intervention was between 1 day and 5 years. In most trials, the insulin pumps were not mainly designed for type 2 diabetes.
In the OpT2mise study (27), a 2-month run-in period with a standardized titration protocol (including basal and bolus adjustments) for dose optimization was scheduled before patients with continued poor glucose control were allocated randomly to the pump group or MDI group. Initiation basal rate for the pump group was 50% of the total daily insulin dose. The significant glycemic improvements observed after 6 months of pump use (HbA1c, −1.1±1.2% vs −0.4±1.1%; P<.0001) also prevailed during the subsequent 6-month follow-up and were reproduced by the MDI group that switched to pump after the first 6 months.
In most trials initiating pump use, a basal insulin delivery of 40%–60% of the total daily insulin dose was chosen. In some studies, patients stopped previous diabetes therapy options when pump therapy was initiated, whereas others used other therapy options like oral diabetes medications. There was a high variation in the extent of education programs the patients underwent during the different trials. Pump therapy requires users and health-care professionals to fully understand how to use the given insulin pump and provided features; therefore, education programs and intensive training providing this knowledge are essential for insulin pump use.
Conclusion
The studies discussed in this article differ considerably regarding study design, study methods, insulin pumps, and study population, making it difficult to compare results directly. Despite being on multiple daily injections, those with type 2 diabetes and high HbA1c levels may benefit from insulin pump therapy. Simpler insulin pumps may be more suited for type 2 diabetes.
Comments
Insulin pumps designed specifically for people with type 2 diabetes include PAQ by CeQur, Simplicity by CeQur (designed for bolus delivery), and V-Go by Valeritas. However, national guidelines do not primarily recommend insulin pump therapy for this patient group. Type 2 diabetes is a heterogeneous disease with substantial phenotypic variability; therefore, it is difficult to provide general conclusions about the effectiveness of pump in type 2 diabetes therapy. The pump types, characteristics, and associated barriers may play a relevant role for therapy outcomes. Authors argue that the pump features and handling for patients with type 2 diabetes should be as easy as possible, indicating a need for insulin pumps specially designed for patients with type 2 diabetes. However, further research is required to determine if pumps designed for type 2 diabetes are more cost-effective than other intensified insulin regimens and which subgroups of patients with type 2 diabetes should be considered for pump therapy. The role of insulin pumps in type 2 diabetes patients is evolving as newer noninsulin agents are introduced, and further randomized studies in those with optimal medical therapy are required.
TECHNOLOGY SAFETY AND ACCEPTANCE
Declining Frequency of Acute Complications Associated with Tubeless Insulin Pump Use: Data from 2,911 Patients in the German/Austrian Diabetes Patienten Verlaufsdokumentation Registry
Biester T1, Schwandt A2,3, Heidtmann B4, Rami-Merhar B5, Haak T6, Festa A7, Kostow S8, Müller A9, Mönkemöller K10, Danne T1; DPV Initiative
1AUF DER BULT Diabetes Center for Children and Adolescents, Hannover, Germany; 2Institute for Epidemiology and Medical Biometry, ZIBMT, Ulm University, Ulm, Germany; 3German Center for Diabetes Research (DZD), Neuherberg, Germany; 4Catholic Children's Hospital Wilhelmstift, Hamburg, Germany; 5Department of Pediatric and Adolescent Medicine, Medical University of Vienna, Vienna, Austria; 6Diabetes Clinic Bad Mergentheim, Bad Mergentheim, Germany; 7Landesklinikum Korneuburg Stockerau, Korneuburg Stockerau, Austria; 8Sana Klinikum Lichtenberg-Oskar-Ziethen-Krankenhaus, Berlin, Germany; 9Klinik für Stoffwechsel und Diabetes, Karlsburg, Germany; 10Kinderklinik Amsterdamerstrasse, Köln, Germany
Background
Severe hypoglycemia and diabetes-related ketoacidosis are life-threatening complications of diabetes. The objective of this paper was to characterize patients with diabetes treated with a tubeless insulin pump (Omnipod Insulin Management System; Insulet Corp, Acton, MA) and to evaluate the frequency of acute complications with long-term use of the system.
Methods
This retrospective analysis of the German/Austrian Diabetes Patienten Verlaufsdokumentation (DPV) registry included data from 3657 patients with diabetes (type 1 diabetes, n=3582; type 2 diabetes, n=25; latent autoimmune diabetes in adults/other, n=50) treated with a tubeless insulin pump. HbA1c levels, frequency of diabetes-related ketoacidosis, and severe hypoglycemia were compared between 1 year before and 1 year (n=2911) or up to 3 years (n=1311) after tubeless insulin pump initiation and compared with a contemporary cohort on multiple daily injections (MDIs) with 3-year data (n=1874). Ketoacidosis was defined as pH <7.3 or bicarbonate concentration <15 mmol/L. Authors counted patients with multiple ketoacidosis events per year as 1. Severe hypoglycemia was evaluated both as (a) Level 3, defined as blood glucose (BG) <70 mg/dL or <3.9 mmol/L and requiring assistance from another person to administer carbohydrates, glucagon, or intravenous glucose actively and (b) coma, defined as loss of consciousness or occurrence of seizures.
Results
A total of 2911 patients had at least 1 year of data using the prior treatment and 1 year of follow-up data after tubeless insulin pump initiation. A subgroup of 1311 patients had 3 years of follow-up data after tubeless insulin pump initiation. The prior treatment modalities for patients with type 1 diabetes were as follows: MDI for 58.5% of patients, another pump for 38.1% of patients, tubeless insulin pump as initial therapy for 3.2% of patients, and unknown for 0.3% of patients. Patients using tubeless insulin pump therapy had a median age of 13.7 years (IQR, 10.8–17.3), diabetes duration of 3.7 years (IQR, 1.7–8.0), and HbA1c 7.5% (IQR, 6.9%–8.2%).
In patients with 3 years of follow-up data (n=1311), 6.3% of patients had ≥1 episode of ketoacidosis, 5.5% had severe hypoglycemia (Level 3, requiring assistance), and 1.7% had severe hypoglycemia (coma) event with prior treatment. After 3 years of tubeless insulin pump therapy, the frequencies of ketoacidosis, severe hypoglycemia (Level 3), and severe hypoglycemia (coma) decreased to 2.2%, 4.1%, and 0.5%, respectively. In the subgroup with 3-year data, ketoacidosis and severe hypoglycemia remained significantly lower compared with MDI after adjustment in multiple regression analysis high rates of treatment adherence. (continuation of pump therapy) (>90%) were observed. The HbA1c was 7.5% with prior therapy, decreased slightly to 7.4% in the first year of tubeless insulin pump therapy, and increased to 7.7% in years 2 and 3.
Conclusions
In this DPV registry study, tubeless insulin pump therapy was associated with reductions in the frequency of ketoacidosis and severe hypoglycemia after 3 years compared to those using multiple injection therapy. However, compared to the MDI group at baseline, the pump group had much higher rates of ketoacidosis (3.1% vs 6.3%, P<.001). Severe hypoglycemia rates were lower after year 2 and year 3 with the tubeless pump group, whereas those rates did not change significantly in the MDI group. HbA1c showed a slight increase in MDI and pump groups (authors ascribed this to age-related deterioration in glycemic control) without an apparent difference between groups (comparison data not reported).
Comments
It is important to highlight that real-world observational data do not prove causality. In this study, baseline ketoacidosis rates of the pump cohort were about double that of MDI patients, and this baseline imbalance may have driven the statistical significance. However, it is reassuring to note that year by year, rates of ketoacidosis in the pump cohort declined while rates remained roughly the same in the comparable MDI group. In contrast, rates of severe hypoglycemia were similar between the groups at baseline, and statistically, a significant reduction was seen in the pump group after 2 and 3 years.
Factors Associated with Insulin Pump Discontinuation in Adults with Diabetes: A Time-to-Invent Analysis and Prediction Model
Gargouri I1, Hadja Inna AA1, Franc S1,2,3, Picaud P2, Penfornis A1,4, Amadou C1,4
1Diabetes Department Sud-Francilien Hôpital, Corbeil-Essonnes, France; 2Center for Study and Research for the Intensification of Diabetes Treatment (CERITD), Bioparc-Génopole, Évry, France; 3LBEPS, Univ Evry, IRBA, Université Paris-Saclay, Évry, France; 4Kremlin Bicetre Medical School, Paris-Saclay University, Paris, France
Background
Reasons for insulin pump discontinuation have been studied in children and adolescents living with diabetes. The authors aimed to determine the probability that adults with diabetes seen at a specialty center would continue pump therapy at 18 months. They additionally sought to determine the factors associated with pump discontinuation and to develop a simple prediction model.
Methods
This retrospective study was conducted at Centre Hospitalier Sud-Francilien and included all adult patients with type 1 or type 2 diabetes who started insulin pump treatment between January 2015 and June 2018. Individuals were excluded for pregnancy, short-term pregnancy plans, and insulin pump discontinuation within the previous 6 months. The probability of insulin pump continuation after 18 months was estimated using the Kaplan-Meier method. Discontinuation was defined as documentation of pump discontinuation in the medical record or two consecutive prescriptions within at least 6 months for multiple daily injection pens. Factors associated with insulin pump discontinuation were studied by using a Cox regression model, and an exponential model was built for prediction purposes.
Results
The study included 315 patients with mean age 41 years and mean duration of diabetes 16 years; 50% were men, 74% had type 1 diabetes, and mean HbA1c of all patients was 9.1% (76 mmol/mol). After 18 months, the rate of insulin pump continuation was 0.80 (95% CI, 0.76–0.85). By multivariate analysis, the occurrence of severe hypoglycemia in the previous year was associated with insulin pump discontinuation (HR 2.42; 95% CI, 1.30–4.51), while other factors did not reach statistical significance. The probability of pump discontinuation was expressed as:
where severe hypoglycemia, previous use of insulin pump, and social deprivation are binary (0 or 1).
Conclusions
Insulin pump discontinuation occurred in 20% of patients at 18 months after initiation and was primarily associated with recent severe hypoglycemia. The type of diabetes and glycemic control at baseline were not associated with treatment discontinuation.
Comments
This single-center retrospective analysis evaluated factors associated with pump discontinuation. The authors make it clear that the proposed exponential multivariate model may only be applicable to their diabetes center. Of note, automated insulin delivery during the evaluated interval consisted only of predictive low-glucose suspend systems. A history of severe hypoglycemia <1 year before pump initiation, social deprivation (unemployment, long-term disability, being a beneficiary of French universal health care or state medical assistance), and any history of prior pump use were found to be predictive of discontinuation. The authors emphasize that all people should be offered diabetes technology, and that these data should not be used to exclude access. Indeed, those likely to discontinue require the most support.
Footnotes
Author Disclosure Statement
Rayhan Lal has research support from the NIDDK (1K23DK122017) and JDRF (2-SRA-2022-1169-M-B). He has consulted for Abbott Diabetes Care, Biolinq, Capillary Biomedical, Deep Valley Labs, Gluroo, Morgan Stanley, Provention Bio, and Tidepool.
Lalantha Leelarathna reports having received speaker honoraria from Animas, Abbott, Insulet, Medtronic, Novo Nordisk, Roche, and Sanofi. Lalantha Leelarathna is on the advisory panels for Animas, Abbott, Novo Nordisk, Dexcom, Medtronic, Sanofi, and Roche. Novo Nordisk and Dexcom provided research support.