A Clinical Overview of Insulin Pump Therapy for the Management of Diabetes: Past, Present, and Future of Intensive Therapy

An insulin pump is a small, digital device that continuously delivers rapid-acting insulin through a small catheter inserted into the subcutaneous tissue and secured in place on the skin with adhesive (referred to as an “infusion set” or “infusion cannula”). In most insulin pumps, the infusion set connects to the pump by plastic tubing, and insulin infuses from the pump through the tubing to the infusion set cannula and into the subcutaneous tissue (Figure 1). Some pumps, referred to as “patch pumps,” do not use tubing and instead adhere directly to the skin. Patch pumps deliver insulin through the infusion cannula and are programmed from a remote device using wireless technology (Figure 2) (24).

Insulin pumps generally use rapid-acting insulin formulations (i.e., insulin lispro, aspart, or glulisine). Lispro and aspart are approved by the U.S. Food and Drug Administration (FDA) for use in a pump insulin reservoir for up to 144 hours, but glulisine should be replaced every 48 hours due to a risk of crystallization. Regular insulin is also FDA-approved for use in pumps and is sometimes used instead of rapid-acting formulations because of its lower cost. Concentrated insulins (e.g., U200 or U500), dilute insulin (e.g., U50 or U10), and ultra-rapid-acting insulin analogs (e.g., Fiasp) are undergoing studies but are not yet FDA-approved for use in pumps.

Insulin pumps deliver insulin in two primary ways: a continuous infusion of rapid-acting insulin throughout the day and night (basal), and discrete, one-time doses of rapid-acting insulin given by the user for meals or high blood glucose correction (bolus). Basal insulin delivery replaces the use of the longer-acting exogenous insulin formulations used in MDI regimens. A multitude of factors influence basal insulin needs, including physiology, developmental life stage (i.e., puberty or growth), activity level, time of day, and sleep schedule. Insulin pumps deliver basal insulin in increments as small as 0.01 unit/hour and permit multiple rates of basal infusion throughout the day and night to best optimize glycemic control and individualize therapy, mimicking nondiabetes physiology. Additionally, many insulin pumps include a temporary basal feature, allowing users to temporarily increase or decrease basal delivery by a percentage relative to the programmed basal rate or by programming a new basal rate. The temporary basal feature is useful for situations in which insulin needs may change drastically, but for a confined period, such as during acute illness or when exercising.

Users can program larger, discrete bolus doses of insulin for carbohydrate consumption and high blood glucose corrections. Most insulin pumps contain a bolus calculator with which the pump calculates a bolus dose recommendation based on the current blood glucose value, current insulin on board (remaining active insulin from previous bolus doses), and total grams of carbohydrates that the user enters into the pump. Some pumps have an extended bolus option, which delivers a portion of the total bolus dose immediately and extends the delivery of the remainder of the dose over a longer period (usually 2–3 hours) to help prevent delayed postprandial hyperglycemia. This option can be helpful when consuming high-fat meals or for individuals with gastroparesis. Pumps deliver bolus doses in increments as small as 0.025 units, allowing for more precise insulin dosing than is possible with insulin pens or syringes. A variety of different insulin pump options are commercially available to individuals with type 1 or type 2 diabetes, and the spectrum of options consistently evolves with new technologies arriving on the market each year (Table 1).

Insulin pump therapy was originally developed for use in type 1 diabetes. However, it may also benefit those with type 2 diabetes who require insulin therapy (7,8). Several studies have demonstrated improved glycemic control for individuals with suboptimally controlled type 2 diabetes treated with multiple oral diabetes medications or an MDI insulin regimen who discontinue all oral medications other than metformin and initiate insulin pump therapy. These studies have reported a reduction in A1C of 1.0% or more with lower total daily insulin requirements, reduced risk of hypoglycemia, and higher treatment satisfaction compared to MDI (25–28). These benefits were obtained using relatively simple insulin dosing regimens. Participants required only one or two basal rates, and use of the bolus calculator to determine bolus doses was not associated with reduction in A1C (28). This result suggests that individuals with type 2 diabetes may not require the complex pump features that are beneficial for the management of type 1 diabetes (29) and that cited barriers to conventional insulin pump therapy for individuals with type 2 diabetes, such as extensive educational requirements and high cost compared to MDI, could be reduced with simpler devices (30).

Thus, simplified pump technology has been developed specifically targeting the needs of people with type 2 diabetes. These novel pumps for type 2 diabetes are small, disposable, patch pumps that adhere to the body with adhesive and consist of an insulin reservoir and an infusion cannula that auto-inserts with the press of a button. The pumps come pre-programmed for basal delivery based on total daily basal dose and permit bolus dosing, via a bolus button, with pre-set bolus delivery increments (i.e., 2-unit insulin delivery per each button press). Currently, these pumps do not have a bolus calculator or any of the other advanced features of conventional insulin pumps.

There are a few patch pumps specifically indicated for individuals with type 2 diabetes; however, only the V-Go (31) is commercially available in the United States. The V-Go is prescribed as V-Go 20, 30, or 40, with the numbers referring to the fixed amount of basal insulin that will be delivered in 24 hours (i.e., V-Go 20 will deliver 20 units of insulin in 24 hours at a single rate of 0.83 units/hour). The V-Go contains a bolus button for meals that permits up to 36 units of bolus insulin delivery per day, in 2-unit increments.

Clinical studies using these novel patch pumps for type 2 diabetes management have demonstrated improve glycemic control, high patient satisfaction, reduced barriers to insulin pump treatment, and cost savings when compared to MDI therapy (32–34).

The development of CGM systems in the early 2000s was followed by the advent of the sensor augmented pump (SAP), which combines a CGM and insulin pump in one system. A CGM device consists of three components: 1) a thin, flexible sensor, which is inserted into the subcutaneous tissue and continuously measures glucose levels in the interstitial fluid, 2) a transmitter that sends the sensor glucose data to a receiver, and 3) a receiver, which displays the glucose values. In an SAP system, the insulin pump pairs to a CGM system and acts as the receiver, displaying CGM sensor glucose data on the pump’s home screen and thus allowing users easy access to the sensor glucose information. SAP systems have shown a greater reduction in A1C (−0.8 ± 0.8% vs. −0.2 ± 0.9%, P <0.001) after 12 months when compared to MDI therapy (35). The STAR 3 study, a randomized controlled trial involving 82 children and adolescents, found that participants using SAP therapy were more likely to meet glycemic targets than those using MDI, and SAP users had reduced glycemic variability after 12 months (36). Those wearing the sensor more consistently in the SAP group were more likely to meet glycemic targets, suggesting that easy access to CGM data on the pump helps individuals respond more readily to high and low glucose values, thus reducing glucose variability (35).

After the advent of SAP, further integration of pumps with CGM devices occurred with hypoglycemia suspension technology. In hypoglycemia suspension systems, the insulin pump not only displays the sensor glucose values, but also automatically suspends insulin delivery in response to hypoglycemia or anticipated hypoglycemia, based on CGM data, in an effort to prevent low blood glucose levels (37). This additional layer of protection against hypoglycemia is vital, as severe hypoglycemia remains the most concerning acute complication of intensive insulin therapy (38,39).

The first insulin pump system with hypoglycemia suspension technology was the Minimed 530G, which suspends insulin delivery when hypoglycemia occurs, a function referred to as “threshold suspend” or “low glucose suspend” (LGS). Studies using LGS have demonstrated a 40–50% reduction in hypoglycemia (<70 mg/dL), without an increase in A1C or mean sensor glucose values compared to SAP therapy alone (40,41). More recently, two more systems have become available in the United States that contain predictive low glucose suspend (PLGS) technology: the Minimed 670G (“suspend before low”) and the t:slim X2 system (Basal-IQ). These systems automatically suspend insulin delivery 30 minutes before hypoglycemia is predicted to occur, based on CGM data. Studies on the effectiveness of PLGS for reducing exposure to nocturnal hypoglycemia have demonstrated a 50–80% reduction in hypoglycemia overnight, without increasing the risk of ketosis (42,43), and an overall 31–50% reduction in hypoglycemia when using PLGS compared to SAP alone, with no increase in mean glucose value or hyperglycemia (44,45).

Hyperglycemia remains a significant challenge in diabetes management, even with the use of an insulin pump. Automated insulin delivery technologies (also referred to as “artificial pancreas” or “closed-loop” systems) aim to reduce hypoglycemia and hyperglycemia, thus improving overall glycemic control and increasing time spent in the glucose target range (70–180 mg/dL). An automated insulin delivery system consists of an insulin pump, a CGM device, and a control algorithm that calculates and dynamically adjusts insulin delivery in real time, based on the CGM sensor glucose values and trends (i.e., as sensor glucose values increase or decrease, insulin delivery increases or decreases as well). The first generation of automated insulin delivery is the hybrid closed-loop (HCL) system, which dynamically modulates basal insulin delivery but still requires users to deliver bolus doses for meals using the bolus calculator. Clinical trials using a variety of automated insulin delivery systems in children and adults have consistently demonstrated improved glycemic control, evidenced by reduction in A1C, increased sensor time in target range, reduced glycemic variability, and reduction in hypoglycemia (46–51).

In 2016, the Minimed 670G became the first automated insulin delivery system to be approved by the FDA for use in children and adults with type 1 diabetes. This system operates in “manual mode” or “auto mode” and pairs with the Guardian 3 CGM device. Manual mode refers to the conventional pump mode, through which basal insulin delivery is dictated by basal rates programmed into the pump. The 670G system in manual mode also contains LGS and PLGS technologies when the pump is paired with CGM. Auto mode refers to the HCL system, through which the pump calculates basal delivery every 5 minutes based on the sensor glucose trends. To use auto mode, the CGM must be active and sensor adequately calibrated. In both modes, the user delivers bolus doses for meals and high blood glucose corrections.

Results from the pivotal trial in adolescents and adults demonstrated safety and effectiveness of the system for a cohort of 30 adolescents and 94 adults (all previous insulin pump users) using the HCL feature (auto mode). A1C decreased by ∼0.5%, and sensor time in range increased by ∼8% after 3 months of use, with no occurrences of severe hypoglycemia or diabetic ketoacidosis (52). Adolescents spent less time in the HCL mode compared to the adults in the trial, indicating that adolescents may have a more difficult time adhering to system requirements to maintain time in the HCL mode (auto mode), such as responding to system alerts and maintaining sensor calibration.

Several other HCL systems are undergoing clinical testing and are expected to become commercially available in the next few years. In addition, the “Do-It-Yourself” (DIY) diabetes community has developed a closed-loop algorithm, used by many individuals with diabetes to essentially build their own closed-loop device. DIY closed-loop systems use commercially available CGM systems and insulin pumps and an open source algorithm run on a smartphone app to automate insulin delivery (53).

Recent clinical guidelines from diabetes organizations worldwide, including the American Diabetes Association, the International Society for Pediatric and Adolescent Diabetes, the Endocrine Society, and the American Association of Clinical Endocrinologists/American College of Endocrinology state that insulin pump therapy may be beneficial for all individuals with type 1 diabetes, regardless of age (54–57). Individuals with type 1 diabetes who are not meeting glycemic targets or have high rates of hypoglycemia or hypoglycemic unawareness may benefit the most from pump therapy. Individuals with gastroparesis may also benefit from pump therapy, specifically from the ability to extend bolus delivery to manage the delayed rise in glucose from meals that occurs with gastroparesis. Finally, even individuals who are meeting their glycemic targets with an MDI regimen but who desire more flexibility in their type 1 diabetes management may find improvements in quality of life and treatment satisfaction when switching to insulin pump therapy. Insulin pump therapy is recommended for individuals with type 2 diabetes who are not meeting glycemic targets with MDI, oral medication, and lifestyle modifications (56).

As advancing insulin pump technologies more fully incorporate CGM, it is important to consider which individuals will benefit from CGM as well when counseling patients on the optimal type of insulin pump therapy for them. SAP systems and automated insulin delivery systems will benefit individuals who can manage CGM and those who are willing to relinquish some control of insulin dosing to the automated pump system. Table 2 provides extensive detail on clinical considerations for different types of insulin pump technologies.

To ensure successful adoption of insulin pump therapies, individuals must be willing to wear an insulin pump and CGM device (when applicable). Additionally, individuals and their caregivers should have appropriate expectations of their insulin pump technology and be able to adhere to the self-care tasks required for success with their chosen technology. This is true for all insulin pump technologies, from the simplest pumps to automated insulin delivery systems. Further, individuals and their caregivers must be cognitively and emotionally able to manage the insulin pump device and solve problems that may arise, such as infusion set malfunctions. Finally, patients must be motivated to complete all of the necessary education on their therapy and follow up regularly with their health care team.

When assessing an individual’s readiness for insulin pump technologies, the individual’s willingness to wear a device on the body and ability to cope with the device’s presence over time are among the most important clinical considerations. In fact, in a study surveying >1,500 adults with type 1 diabetes, one of the most commonly endorsed barriers to device uptake was the hassle of wearing a device and disliking having devices on one’s body (58). Further, a recent review summarizing biopsychosocial factors related to sustained device use reported that body image concerns are a major barrier to device use for both adolescents and adults (59). Individuals have reported feeling self-conscious during intimacy or feeling like a “cyborg”’ being “shackled” to multiple devices (58,60–62). Clinicians should discuss these concerns with their patients regularly and assist patients in increasing their problem-solving capacity and social support to reduce these barriers and facilitate sustained device use. Ultimately, insulin pump and CGM manufacturers need to continue working to reduce the size and improve the discreetness of insulin pumps and CGM systems to increase the uptake of diabetes devices and reduce the risk of discontinuation among patients who try them (63,64).

In addition to the psychosocial concerns about wearing devices, there are also numerous practical issues involved in wearing devices. Many people struggle to keep devices adhered to the body, which is especially true for young children, individuals who participate in sports or other physical activities, and individuals who experience heavy sweating independent of activity level. Further, adverse skin reactions to infusion set and CGM adhesives are a common reason for discontinuation of insulin pump and CGM therapies. Clinicians should assess skin integrity and tolerance to adhesives and work with their patients to overcome these barriers (65). Many products are available to help protect the skin from irritation, such as barrier films and hypoallergenic over tapes; such products may help people keep their devices in place while also reducing the incidence of skin reactions (65). Finally, clinicians should educate their patients on the importance of site selection and site rotation to avoid other skin issues with chronic device wear, such as lipohypertrophy.

One of the most important roles clinicians play in optimizing their patients’ success with insulin pump therapy is setting appropriate expectations for the devices. Clinicians should assess their patients’ expectations of the therapy, including why they desire to use a particular insulin pump, what they expect the system to be like, and what type of self-care they think is required of the user for the device to operate properly. A balanced discussion of the potential benefits and drawbacks of the preferred system should occur. Unrealistic expectations of any diabetes technology increase the risk for dissatisfaction, suboptimal glycemic control, and discontinuation of device use (60,61,66–68). Individuals with diabetes and their caregivers must understand the limitations of insulin pump technologies and the potential problems they may encounter, such as infusion set failure, pump malfunction, skin irritation, and alarm fatigue. Use of CGM requires responding to alarms and managing difficulties such as lost sensor signals or errors in calibration. Further, it is imperative that individuals conceptualize insulin pump technologies as a tool to help them improve their diabetes management and not as a panacea that will “cure” their diabetes or eliminate the need for self-care.

Expectation-setting is even more important for automated insulin delivery systems (69). Many individuals with diabetes expect these systems to take over their diabetes care for them (64), and, to date, this is not a realistic expectation. Early HCL systems available today, such as the Minimed 670G, may require users to check blood glucose levels, calibrate the CGM device, count carbohydrates, administer boluses for meals, and respond to system alerts.

Performing self-care behaviors is especially important for glycemic control and safety while using an insulin pump. Several studies have shown that missed meal boluses (70–72) and lack of correction boluses (73) in response to hyperglycemia predict suboptimal glycemic control, especially for adolescents with type 1 diabetes. Further, consistent monitoring of glucose values is highly correlated to improved glycemic control (74) and remains an important behavior for success with all insulin pump technologies. It is important that clinicians communicate clearly about the importance of adherence to self-care with insulin pump therapy and work with individuals to overcome any barriers to self-care.

Initiating insulin pump therapy requires extensive education and frequent follow-up with the health care team. Education should be ongoing and individualized to teach advanced skills over time (i.e., the use of extended boluses, hypoglycemia suspension, and HCL features), based on individuals’ specific diabetes management needs. Individuals with diabetes should be encouraged to complete the necessary education, and their families/caregivers must also be integrated into the educational process. Further, health care providers must be able to provide their patients with expert education on using insulin pump therapy and adequate clinical follow-up.

The American Association of Diabetes Educators provides guidance and practical tips on how to assess individual readiness for pump therapy, educate individuals on the basics of insulin pump therapy, and provide sufficient follow-up to support success (75). All pump education programs should include instruction on basic pump operation, including inserting the infusion set, changing the insulin reservoir, programming insulin pump settings, and delivering boluses. Initial pump education should also include device troubleshooting and guidance on managing persistent hyperglycemia, including when and how to check ketone levels, give a subcutaneous injection, and change the infusion set (76). Finally, individuals need frequent contact with their health care team in the initial weeks to optimize basal and bolus insulin pump settings.

When initiating advanced insulin pump therapies such as SAP or automated insulin delivery systems, the educational approach must be individualized. For those new to CGM or insulin pump therapy or those who struggled using either technology in the past, initiating both the pump and CGM at once may be overwhelming. Likewise, ensuring success and confidence in the basics of each technology is paramount before adding advanced features such as hypoglycemia suspension or automated insulin delivery. Thus, completing education on each component of a system separately, before integrating the technologies, may increase success with advancing therapies (77).

There is no one-size-fits-all approach to insulin pump therapy, and fortunately, there are many options for clinicians to consider with each patient with diabetes. Insulin pump technologies are advancing at an extraordinary rate and have potential to improve diabetes outcomes for individuals of all ages with type 1 or type 2 diabetes. However, individuals with diabetes must be able to overcome any barriers to device wear, have realistic expectations of their particular device, perform self-care, and complete extensive education and clinical follow-up to realize success with insulin pump therapies. It is important for clinicians to work with individuals with diabetes and their caregivers to optimize use of insulin pump technologies initially and into the future.

C.B. is a contracted product trainer for Medtronic and has received speaking honoraria from Insulet. L.H.M. is a contracted product trainer for Medtronic and a consultant for Tandem Diabetes, Clinical Sensors, and Capillary Biomedical. G.P.F. receives research support from Abbott, Beta Bionics, Dexcom, Insulet, Medtronic, Tandem, and Type Zero. He has served as a speaker/consultant for Dexcom, Medtronic, and Tandem. No other potential conflicts of interest relevant to this article were reported.

C.B. wrote the manuscript and conducted the literature review, L.H.M. and G.P.F. reviewed/edited the manuscript. C.B. is the guarantor of this work, and, as such, takes responsibility for the integrity of the manuscript and the accuracy of the information presented.