Quo vadis diabetes technology?

Summary

Management of patients with diabetes mellitus who require intensive insulin therapy has changed dramatically. From urine and then whole blood glucose monitoring patients can now wear personal continuous glucose monitors; from using multiple insulin injections they can now deliver subcutaneous insulin continuously using ever-more sophisticated insulin pumps. With the current integration of continuous glucose monitoring and insulin pumps a new era is arriving when “smart” insulin pumps can automate insulin delivery to assure more stable glycemic control while minimizing episodes of hypoglycemia. This review briefly summarizes results of key research studies demonstrating feasibility of the technology. It also discusses patient selection and potential challenges faced by attempting to incorporate diabetes technology into daily clinical practice. Finally, an overview of integrated systems in current and planned clinical trials is presented.

Keywords:
diabetes technology, insulin pumps, continuous glucose monitoring, automated insulin delivery, patient education

INTRODUCTION

With the recent breakneck speed with which diabetes technology has advanced, one can indeed ask: “Where are you going?” Can this progress advance and improve care of our patients with diabetes? If so, how and when? And will every patient with diabetes mellitus benefit? If it is to applied widely, who will teach the clinicians and patients clinical applications of this technology? And, who will pay for the education efforts and the technology itself? What data are needed to persuade the payers (insurance companies, governments, employers and other purchasers of health insurance, and patients themselves)? We can ask many such relevant questions. before diving more deeply into the subject matter. And, in a way, it is a good problem to have, to be able even to contemplate there is a perhaps better and safer way to active the glycemic goals for our patients with diabetes. One would hope that these efforts would eventually result in prevention or delay of those dreaded and expensive long-term diabetic complications. It is those, rather than the current hyperglycemic levels, which cost us all (as patients, relatives, taxpayers) dearly in terms of quality of life, longevity and money.

DIABETES COMPENSATION IN PATIENTS TREATED WITH AN INSULIN PUMP

The opportunity to more closely monitor, and then respond to, changing glucose levels in real time in people living with diabetes offers the potential for improving glucose control safely. Continuous glucose monitoring (CGM) devices generally offer real-time display and features that allow patients to respond to changing glucose levels as they occur. In real-world settings, however, patients may not understand, interpret, and respond appropriately to data they view. This presents an important challenge for integrating insulin infusion pumps and CGM into practice for patients with type 1 diabetes mellitus (T1DM) and type 2 diabetes mellitus (T2DM). Adoption of these technologies has been slow: currently, CGM is used by approximately 15% of people with T1DM and fewer with T2DM [1].

Previous results (STAR 3 [Sensor-Augmented Pump Therapy for A1C Reduction phase 3 trial], 2012 [2]) demonstrated improved glucose control with a system of combined continuous subcutaneous insulin infusion (CSII) pump and a CGM system (sensor-augmented pump) therapy in patients with T1D. This system includes threshold-suspend functionality [3], which may increase the performance and frequency of CGM use in adult and pediatric patients.

Newer developments (such as in the marketed predictive low-glucose suspend as in MiniMed 640G system or hybrid closed loop option in MiniMed 670G system) offer evidence of further progress, allowing for even greater patient safety, by suspending insulin infusion before hypoglycemia occurs. [4]. A meta-analysis in 2012 of 10 trials supported the superiority of CGM over self-monitoring of blood glucose (SMBG) and sensor-augmented pump devices over multiple daily injections (MDI) plus SMBG in terms of glycated hemoglobin (hemoglobin A1C, or A1C) reduction, without increased risk of hypoglycemia [5]. A consensus statement in the same year [6] recommended clinicians consider CGM in children and adolescents with T1DM, particularly the following patients:

• performing frequent blood glucose monitoring,
• repeated and/or severe hypoglycemic episodes,
• hypoglycemia unawareness, especially in young children,
• nocturnal hypoglycemia,
• wide glucose excursions, regardless of A1C,
• suboptimal glycemic control, with A1C exceeding the target range,
• tight glycemic control, A1C levels <7%, in patients who wish to maintain target glycemic control while limiting hypoglycemia risk.

American Association of Clinical Endocrinologists (AACE) recommends that CGM be considered for all insulin-using patients, regardless of diabetes type given the ever-present risk of hypoglycemia [7]. Insulin pump usage is, of course individual patient’s choice but is recommended for consideration in patients already undergoing intensive insulin therapy whether for T1DM or T2DM (i.e for those patients who currently perform at least 4 insulin injections and 4 SMBG measurements daily) [8]. Integration of CGM and CSII may be considered in patients already on CSII or who are appropriate for initiating CSII.

Several devices that integrate insulin infusion pumps and CGM currently on the US market include the following:

• Medtronic/MiniMed insulin pump/integrated systems (MiniMed 630G and 670G insulin pump systems, MiniMed 530G with Enlite sensor, MiniMed Paradigm^® REAL-Time Revel^TM System) and their proprietary CGM sensors (iPRO, including the iPRO2 professional CGM sensor, Enlite, and Guardian^® Sensor 3) [9];
• Tandem’s t:slim X2™ insulin pump integrated with Dexcom CGM sensor (either G4^® Platinum or G5^® Mobile CGM) [10];
• Animas^® Vibe^® system integrated with Dexcom CGM sensor (G4^® Platinum CGM) [11];

Medtronic/MiniMed pumps from 530G onward have carried integration to the next step with a threshold-suspend feature that automatically holds insulin delivery for up to 2 hours once blood glucose level decreases below the threshold, if a patient does not respond to an alert/alarm. MiniMed 670G is a hybrid closed-loop system that automatically integrates CGM and insulin pump delivery striving to maintain glucose level at 120 mg/dL overnight and between meals. There are no longer fixed basal rates; rather the “brain” inside the pump directs the device to send insulin boluses and when needed to maintain the pre-set glucose value. Patients, however, must still enter the carbohydrate content before meals to deliver the necessary meal bolus [9]. Recent results suggest that closed-loop control with CGM and insulin pump improves glycemic control [12,13].

Personal CGM should ideally be considered in all patients with T1DM, in particular those with a history of severe hypoglycemia, hypoglycemia unawareness and to assist in the correction of hyperglycemia in patients not at goal [7,14-17]. Of note, usage and persistence of usage of CGM is lower in pediatric patients [18,19]. The benefits of CGM in patients with T2DM have not been investigated to the same degree [20,21]. For patients with T2DM, in addition to those on intensive insulin therapy one could consider CGM for those on basal insulin or sulfonylurea therapy, and especially in those patients with hypoglycemia unawareness, elderly, living alone, or with advanced renal disease, i.e those who would be considered more vulnerable. Use in this setting is based on individual clinical judgment, as data continue to emerge. CGM is not recommended in patients at low risk of hypoglycemia (those on lifestyle modification alone or on non-hypoglycemic medications). Of note, CGM may provide useful information to adjust insulin dosing and serve as a teaching tool in women with gestational diabetes [22,23].

A NEED FOR CAUTION

While offering the advantages to more closely monitor glucose and to assure safety of insulin delivery, CGM technology still has limitations, notably variable accuracy in the first hours of sensor use, lead-lag phenomenon with rapid glucose changes, and larger absolute relative differences in the hypoglycemic range [7,24-26]. The accuracy of these devices has been rapidly improving. When expressed in terms of Mean Absolute Relative Difference (MARD), for example, ten years ago Dexcom STS system (which had 3-day duration) had MARD of 26% while the current G5 Mobile sensor (with its 7-day patient use) has MARD about 9%. For comparison, for finger stick capillary glucose monitoring, the ISO 2013 require that 95% of blood glucose results should be: within ± 0.83 mmol/L of laboratory results at concentrations of under 5.6 mmol/L (within ± 15 mg/dl of laboratory results at concentrations of under 100 mg/dL); within ± 20% of laboratory results at concentrations of 5.6 mmol/L (100 mg/dL) or more.

Advances in CGM and pump technology have also made use of these devices more complex. Some patients have difficulty knowing how to respond to unexplained hyper- or hypoglycemic events or unexpected pump and/or CGM error messages or alarms. These also present diagnostic and therapeutic challenges to physicians and other clinicians involved in diabetes management. Occurrences such as these could pose serious clinical safety concerns in patients who have not been carefully selected for CGM and CSII or who have not been adequately trained [8]. Proper patient selection and adequate patient education and training in using devices are critical.

Certain comorbid conditions may increase risks associated with intensive insulin therapy, whether delivered by CSII or MDI. Chronic renal failure serves as a prime example of such a condition. In addition, patient mental status (for example, due to depression, mood disorders, cognitive dysfunction, and excessive sedation due to medication) may affect ability to safely use an insulin pump [8].

IDENTIFYING APPROPRIATE CANDIDATES

Appropriate candidates for pump therapy include:

• patients with T1DM who are not at glycemic goal, despite adherence to maximum MDI, in particular:
• those with erratic and wide glycemic excursions (including recurrent DKA),
• frequent severe hypoglycemia and/or hypoglycemia unawareness,
• significant "dawn phenomenon," extreme insulin sensitivity;
• T1DM special populations (including preconception, pregnancy, children, adolescents, and competitive athletes);
• patients with T1DM who feel CSII would help achieve and maintain glycemic targets;
• select patients with insulin-requiring T2DM with any or all of the following:
• C-peptide positivity with suboptimal control on maximal basal/bolus injections,
• substantial "dawn phenomenon",
• erratic lifestyle (e.g. frequent long-distance travel, shift work, and unpredictable schedules),
• severe insulin resistance;
• Select patients with other DM types (e.g. postpancreatectomy) [8].

Importantly, patients who are unable or unwilling to perform MDI, frequent SMBG, and carbohydrate counting, lack motivation to achieve tighter glucose control or have a history of nonadherence, have a history of serious psychological or psychiatric conditions, or have either substantial reservations or unrealistic expectations about pump therapy are not good candidates for pump therapy [8].

All patients who are candidates for CSII are also, by definition, already candidates for CGM. CGM is recommended in all insulin-using patients, regardless of diabetes type, as well as patients with hypoglycemia unawareness and those at risk of hypoglycemia (e.g. elderly patients, those with renal impairment, and athletes) [7].

Use of CGM with integrated pump requires self-management from patients. The ideal candidate must be willing and able to carry out tasks associated with using the system, self-monitor and react to data collected, and maintain frequent contact with the health care team [8]. Intensive education is needed and patients must be willing to complete the necessary training. Family support, particularly with pediatric patients, is paramount to success. The increased burden on patients and their families, as well as health-economic and ethical concerns, must be considered carefully, and this strategy may not be ideal for all patients [27,28].

PHYSICIAN AND PATIENT EDUCATION

Optimal use of CSII and CGM technology depends on clinicians and patients understanding the collection and interpretation of collected information and then acting on that data in a meaningful way. Merely monitoring and recording glycemic values without a plan for acting on those data will not result in better management.

All CGM users must receive education to understand sensor insertion, calibration, and real-time data interpretation. Pediatric users and their caretakers may need additional education, including more in-depth training as well as more frequent follow-up, to successfully adopt this technology [25]. Adoption of standardized data sets in all downloadable CGM reports may enable improved understanding of these data, allowing both patients and physicians to become more confident in using them in daily practice. Wider adoption of CGM in the clinical setting [7] may, in turn, lead to improved glucose management resulting in a lower risk for acute and chronic diabetic complications.

Patients must also receive education on the meaning of pump alerts and alarms, on the need for backup supplies in case of pump or infusion set failure, on adjustments of pump settings, and on handling of various emergency situations. Periodic re-education and retraining is necessary, and new education and training is needed any time patients change or upgrade devices [8].

INTEGRATING INFUSION PUMPS AND CGM

Studies of sensor-augmented pumps have shown consistent improvements in clinical outcomes, including increased rates of users achieving glycemic target, shorter time to achieving target, increased time within target range, lower incidences of hypo- and hyperglycemia, and reduced incidence of nocturnal hypoglycemia (see Table for summary of key results). [2, 4, 13, 29, 31–33]. It is expected that integrated CSII/CGM devices, by improving glycemic control and reducing hypoglycemic event, will improve life expectancy and delay diabetes-related complications in people with T1DM, compared with MDI [31, 34].

Advances in integrated infusion pump and CGM technology has now led to fully automated insulin delivery systems. Results of a randomized, crossover study showed this strategy improved glycemic control in the home setting [29]. A meta-analysis revealed that automated insulin delivery ("artificial pancreas") systems uniformly improved glucose control in patients with T1DM in outpatient settings [30], suggesting this approach may be increasingly important in the real-world setting. In addition, a randomized, crossover trial showed that use of bihormonal (insulin and glucagon) “bionic pancreas” improved glycemic control and reduced hypoglycemia over conventional insulin pump therapy in pediatric patients with T1DM [33].

A review of registries and databases comparing two newer devices with MDI or CSII plus either SMBG or CGM showed that both integrated and stand-alone systems were effective in maintaining glycemic control and reducing hypoglycemic events. Currently, integrated systems may be less cost-effective, as stand-alone systems are initially less expensive but appear to be no less effective [35].

Closed loop, or automated insulin delivery, systems consistently demonstrate improved glucose control. A systematic review showed uniformly improved glucose control in the outpatient setting [30]. A randomized, crossover study of a closed-loop system overnight versus a sensor-augmented pump with low-glucose suspend function in adults and adolescents showed that both approaches were effective, with no statistically significant difference in primary outcome (time within target range) for the whole group. However, among adults, use of the closed-loop system improved time within target range, and reduced both the time above and below target and reduced glycemic variability. Further, it increased treatment satisfaction [32].

Economic analyses have shown added costs associated with integrated CSII and CGM, but with improved quality of life and life expectancy, as well as lower rates of complications. Analyses in the UK, Sweden, and France showed that use of a suspend function may be cost-effective compared with CSII plus SMBG in patients with T1DM [36–38] although there has been no comparable analysis for the US as of yet.

The ideal approach to integrate CSII and CGM in the clinical setting is still under investigation. Data from a randomized trial suggest that use of CGM prior to initiating CSII may improve glycemic control and patient satisfaction in patients with T1DM using sensor-augmented pumps [39]. A key aspect of successful glycemic control with CGM, however, is patients' ability to understand and respond to the data they receive in real time. Recent results show there is variation in how patients adjust insulin therapy. Nonetheless, CGM users do rely on glucose rate of change arrows on their devices to adjust insulin delivery [40].

The integration of insulin pumps and CGM is a rapidly developing area with several upgrades of existing systems and new systems likely to enter the fray (see Table for an overview of the pipeline). Only time will tell which products, if any, are able to produce data of satisfactory quality to convince the regulators, clinicians and ultimately the patients that they are worth the investment of time and money. The major challenge in this field of technology is both the automation of as many processes as feasible and convincing the consumers that their use will bring significant improvements in their quality of life.

Abbreviations:

A1C glycated hemoglobin

BGM blood glucose monitoring

CE European conformity

CSII continuous subcutaneous insulin infusion

CGM continuous glucose monitoring

MDI multiple daily injections

SAP sensor-augmented pump

SMBG self-monitoring of blood glucose

STAR3 Sensor-Augmented Pump Therapy for A1C Reduction phase 3 trial

T1DM type 1 diabetes mellitus

T2DM type 2 diabetes mellitus

George Grunberger

Chairman, Grunberger Diabetes Institute (Bloomfield Hills, Michigan, USA)

Clinical Professor, Internal Medicine and Molecular Medicine & Genetics, Wayne State University School of Medicine (Detroit, Michigan, USA)

Professor, Internal Medicine, Oakland University William Beaumont School of Medicine (Rochester, Michigan, USA)

Visiting Professor, Internal Medicine, First Faculty of Medicine, Charles University (Prague, Czech Republic)

Past President, American Association of Clinical Endocrinologists

President Elect, American College of Endocrinology

e-mail: grunberger@gdi-pc.com