Artificial Pancreas Systems and Physical Activity in Patients with Type 1 Diabetes: Challenges,Adopted Approaches,and Future Perspectives

Abstract

Physical activity is important for patients living with type 1 diabetes (T1D) but limited by the challenges associated with physical activity induced glucose variability. Optimizing glycemic control without increasing the risk of hypoglycemia is still a hurdle despite many advances in insulin formulations, delivery methods, and continuous glucose monitoring systems. In this respect, the artificial pancreas (AP) system is a promising therapeutic option for a safer practice of physical activity in the context of T1D. It is important that healthcare professionals as well as patients acquire the necessary knowledge about how the AP system works, its limits, and how glucose control is regulated during physical activity. This review aims to examine the current state of knowledge on exercise-related glucose variations especially hypoglycemic risk in T1D and to discuss their effects on the use and development of AP systems. Though effective and highly promising, these systems warrant further research for an optimized use around exercise.

Keywords

type 1 diabetes physical exercise hypoglycemia artificial pancreas

Introduction

Patients living with type 1 diabetes (T1D) are dependent on insulin replacement due to autoimmune destruction of β-cells. Additionally, they experience dysregulation in the response of other glucoregulatory hormones (e.g., glucagon and epinephrine), making glucose control challenging especially during exercise.¹

Exercise can be beneficial for patients with T1D with known improvements in cardiovascular fitness, vascular health, insulin requirements, body composition, and quality of life.² However, mainly because of the inability to reduce insulin at the time of exercise, exercise-induced hypoglycemia is frequent^1,3-5 and remains the main limitation for exercise practice.^6-9 Guidelines for minimizing exercise-related hypo- and hyperglycemic risks exist but their implementation require major individualization^1,10,11 as numerous factors influence glycemia during exercise such as the type, intensity and duration of exercise, the timing in relation to meals, and glucose level at exercise start and should be taken into account to adjust insulin dosing and carbohydrate (CHO) intake.¹² Because glucose fluctuations can occur during as well as for numerous hours after exercise (e.g., nocturnal hypoglycemia), intensive blood glucose (BG) monitoring is required to implement two major therapeutic actions: insulin dose modifications (mainly anticipated reduction) and/or additional CHO consumption.^4,9,13 However, exercise reduces the accuracy of continuous glucose monitoring (CGM) systems further complexifying therapeutic decisions.^14-18

Technological strategies aim to reproduce endogenous hormone secretion pattern of the healthy pancreas, decreasing hyperglycemia without increasing the risk of hypoglycemia. Currently, the artificial pancreas (AP) system is the most advanced technology to help reduce the risk of exercise-induced hypoglycemia.³

The objectives of this review are (1) to summarize the current state of knowledge on physiological mechanisms of BG variability and known strategies to prevent hypo- and hyperglycemia around exercise, (2) to discuss how advances in the AP technology could be used to alleviate the burden of glucose management around exercise, and (3) to discuss remaining challenges to optimize this technology.

Physical Exercise: A Challenge for Patients Living With T1D

Glucose regulation will mainly depend on the balance between muscle glucose uptake versus hepatic glucose production and/or ingested glucose. Skeletal muscle uses glucose (glycogen stores and plasma glucose) as a source of energy during exercise and represents the major site for insulin-stimulated glucose uptake (Figure 1).¹⁹ The main hormones implicated in glucose regulation are insulin (muscle glucose utilization and hepatic glucose production inhibition) and glucagon (hepatic glucose production) with the additional role of catecholamines hepatic glucose production in the cases of very intense exercise.²⁰

Figure 1.

Underlying mechanisms inducing hypoglycemia in individuals with type 1 diabetes during low-to-moderate intensity exercise level exercise.

During exercise, in healthy individuals, the combination of rapid insulin secretion suppression with increased counter-regulatory hormonal activation (e.g., glucagon, catecholamines, growth hormone, and cortisol) increases hepatic glucose production to match peripheral (muscle) glucose disposal resulting in minimal glucose varaitions.^9,21 However, patients with T1D are unable to suppress injected exogenous insulin and also frequently to increase counter-regulatory hormonal increasing hypoglycemic risk.¹ Conversely, if insulin levels are too low (e.g., high glucose before exercise) and/or in the context of large catecholamine release (e.g., very intense exercise), hepatic glucose production will exceed muscle glucose utilization resulting in hyperglycemia.^22,23

Exercise intensity is a major factor impacting BG regulation.²⁴ While low and moderate intensity exercise (e.g., jogging, running, the elliptical, swimming, and cycling) are more frequently associated with a hypoglycemic risk, very intense exercise (e.g., strength training, sprinting, interval training, circuit training, and some competitive sports) implies a dominant role for catecholamines with the increase in hepatic glucose production that could lead to hyperglycemia.^22,23 Such hyperglycemia can persist during recovery.^25-27

In addition, an exercise, even at a low intensity, will increase glucose disposal via insulin-mediated and noninsulin-mediated mechanisms for numerous hours especially at the time glycogen stores will be replenished explaining that exercise can also be associated with the risk of delayed, mainly nocturnal, hypoglycemia.^3,28

Classical Strategies for Managing Hypoglycemia During Exercise

Strategies for minimizing hypoglycemic risk around exercise exist but it remains challenging to apply to a specific type, intensity, duration, timing, etc., exercise session^29,30,31: the simplest strategy to apply is CHO supplements,^32,33 especially for longer exercise, unanticipated exercise, and/or during postabsorptive periods (greater than or equal to three hours following meals). However, this strategy may not be optimal/desired by patients if weight loss is a goal of the exercise practice.¹ In addition, key aspects have not yet been well studied such as required CHO amount or optimal intake timing, and thus, it is hard to give precise advice to patients. Another strategy is adding high-intensity sprints in intervals or resistance training at beginning or end of exercise.¹⁰ This approach takes advantage of the counter-regulatory hormones, especially the catecholamines that are secreted during high-intensity exercise. Also, we use a premeal bolus insulin-dose reduction^4,34 for both multiple daily injections and continuous subcutaneous insulin infusion (CSII), when undertaken in the postprandial period (two to three hours postmeal). Recommendation for premeal bolus reduction is proportional to exercise duration and intensity: larger for longer duration and at higher intensity.³⁵ It should be noted that this approach can lead to postexercise hyperglycemic trend. For patients using CSII temporary basal rate, insulin reduction is recommended but this needs to be implemented more than 40 minutes before exercise to get a significant effect.^3,5,36 For more security, many patients with T1D use sensor-augmented pump (SAP) therapy, which allows insulin delivery to be suspended automatically for up to two hours when the CGM detects a glucose level that has reached a prespecified threshold or, in more recent version, a prediction of an imminent hypoglycemia risk. This technology is the first step along the way to closing the loop of glucose control with the AP systems. However, even when one or combinations of these strategies are used, most patients still have wide exercise-related glucose fluctuations.

Artificial Pancreas Systems During Exercise: Evolution and Challenges

Artificial pancreas systems will automate CSII basal rate delivery by using a control algorithm that receives information from interstitial glucose sensors. It computes the optimal hormones (insulin: AP systems with single-hormone “SHAP” or insulin and glucagon: AP systems with dual-hormone “DHAP”) infusion rate to maintain glucose levels within a predefined target range, thus minimizing hypo- and hyperglycemia episodes³⁷ (Figure 2).

Figure 2.

The three components of an artificial pancreas: glucose sensor, control system, and insulin pump.

In recent decades, there has been impressive advancement with AP systems technology to improve diabetes management.^38,11 Device developments have progressed to such a degree that AP systems show improvements in time-spent-in-target BG levels compared to CSII systems when tested in small sample groups under free-living conditions.³⁹ A recent meta-analysis³⁹ indicated that, compared to SAP, a daily mean three hours time-in-target range improvement with 2.5 hours less in hyperglycemic and 30 minutes less in hypoglycemic range.

Despite these advances, perfecting the system to deal with situations that induce a rapid change in BG (e.g., during exercise and/or postmeal) is still a challenge.^21,40-43 Moreover, it is important to note one limitation when reviewing the efficacy of AP studies during exercise. Artificial pancreas algorithms in different clinical trials are based on distinct designs which dictate how the glucose is controlled. The most commonly used ones are based on model predictive control, proportional-integral-derivative control, and fuzzy logic control.^44,45 Technical differences among AP algorithms are beyond the scope of this review, but in the future with enough accumulating data collected during exercise with each controller type, clearer comparisons might become possible.

Overview of the AP Systems and Challenges

As detailed above, patients with T1D are faced with an arduous process to achieve optimal glucose control (minimally glucose fluctuations) and this is even more pronounced during exercise. Likewise, the automation of these decisions with an AP system will ideally necessitate timely and accurate information about these factors in association with exercise.⁴⁰

The variability in exercise modalities is the most important challenge to creating a system that will be able to adapt to all types and duration of activity for a wide variety of individuals, with a broad range of BG levels at the start of exercise and then response during exercise. The design of a complex sensor that recognizes not only changes in BG but also the start, type, intensity, and end of the exercise would be paramount.^39,40

There are additional limitations imposed by the components of AP systems, namely CGM accuracy and rapid insulin analogs pharmacokinetics. (1) Continuous glucose monitoring accuracy is dependent on several factors during exercise such as the physiological delay needed to equilibrate glucose levels between blood and interstitial compartments, intrinsic sensor delays (depending on sensor type) as well as changes in body temperature and blood flow around sensor insertion site.^46-48 Continuous glucose monitoring systems have many algorithms to adjust for these factors in addition to filtering signal noises. While CGM accuracy has significantly improved with newer sensor generations, it is not yet optimal. This often translates into lower a mean tendency toward overestimation of actual BG levels when dropping rapidly or underestimation with fast rising levels depending on the type of exercise.^14-18 (2) When CGM accuracy is added to the pharmacokinetics of available subcutaneous rapid insulin analogs, a large inertia is created in the system making automated glucose control during exercise extremely challenging: during moderate exercise, the previously injected insulin will continue to be absorbed from subcutaneous depots increasing hypoglycemia risks. On the other hand, when BG levels are increasing rapidly during high intensity exercise, additional insulin boluses or dosing might be needed but face delays to be absorbed and to exert an effect in a timely manner.^1,3 Therefore, additional algorithms that can detect and discriminate the type of practiced exercise to predict direction and extent of glucose changes might be needed for optimal functioning of AP systems.^49,50 Low-to-moderate intensity exercise is the most frequently practiced type of exercise by patients living with T1D and was thus included in most of the conducted studies testing the efficacy of AP systems during exercise with only few reports tackling high-intensity sessions.

A fully automated AP without additional inputs was tested by Elleri et al.⁵¹ In this study, 12-pediatric subjects used a SHAP for 36 hours that included 40 minutes morning and 20 minutes evening moderate-intensity exercise sessions. Unfortunately, SHAP did not eliminate the risk of exercise-induced hypoglycemia, and five hypoglycemic episodes (<54 mg/dL) were recorded. Comparable results were found in adult individuals with T1D.⁵²

Thus, with the current rapid acting insulins and glucose sensors, an AP system that is based solely on CGM measurements probably cannot act fast enough to minimize exercise-induced hypoglycemia.⁵³

To improve AP systems performance around exercise, researchers use four types of approaches: (1) automated exercise detection or manual exercise announcement, (2) dual-hormone closed-loop strategy, (3) CHO supplementation, and (4) combinations of two or more of these approaches (Figure 3).

Figure 3.

Challenges and solutions related to the use of the artificial pancreas systems during exercise.

Automated Exercise Detection Strategy

Biometric variables such as heart rate, near-body temperature, skin temperature, energy expenditure, heat flux, accelerometry, and galvanic skin response have been shown to correlate with changes in BG during various types of exercise.⁴⁹ Clinical outcomes of this study indicated that skin temperature, heart rate, and energy expenditure were the best variables to provide information that could help the AP systems to improve glucose levels during different exercise intensities.⁴⁹

In a recent experiment, Turkosy et al⁵⁴ challenged their multimodule multivariable (inputs: energy expenditure, galvanic skin response, and the subcutaneous glucose value from the CGM) adaptive SHAP under various types and intensities of exercise (HIIT, high-intensity interval training; MCT, moderate continuous training; and RT, resistance training) without requiring any manual information to be entered by the user. The authors note that, with their AP systems, regardless of the intensity of the exercise, there is a consistent decline in BG in both exercise and recovery period. The percent of time spent in significant hypoglycemia (<55 mg/dL) was low despite the lack of any manual exercise announcement (%time in hypoglycemia: HIIT, 0.0% ± 0.0%; MCT, 1.73% ± 5.03%; and RT, 1.03% ± 4.37%). The main limitation of this study is the absence of a control arm, so the performance of their AP systems cannot be compared with any standard of care (e.g., CSII).

To reduce the burden on patients with T1D, Jacobs et al⁵⁵ added a heart rate monitor to their SHAP to detect exercise (45 minutes at 60% maximal heart rate) automatically without requiring any manual information to be entered by the user. When heart rate exceeded 125% of resting levels, this triggered a change in glucose target from 111 to140 mg/dL and subsequent insulin infusion by the AP systems. A decrease in time-spent-in-hypoglycemia (<70 mg/dL) was significantly reduced with heart rate signal integration to SHAP 0.5% ± 2.1% vs 7.4% ± 12.5% (P = .028) without an effect on the incidence of hypoglycemic events which was, however, low in both arms.

The use of exercise detection that will use additional AP system inputs may be beneficial in the anticipation of hypoglycemia risk. However, any AP systems that are developed using current CGM and CSII methods will be limited by slow pharmacokinetics of insulin and lagging CGM glucose response when BG declines rapidly. Thus, turning off/reducing insulin infusion upon sensing the start of exercise may be insufficient to completely prevent hypoglycemia,⁵ except maybe if exercise is performed shortly after a meal.⁵⁶

Dual-hormone (Insulin and Glucagon) AP Systems Strategy

The addition of glucagon to AP systems can be used in two ways: (1) preventive approach where glucagon mini-boluses are recommended when total insulin infusion suspension is not enough, aiming to prevent hypoglycemia and (2) the AP systems recommend more important/aggressive insulin rates aiming for lower mean BG with occasional glucagon delivery to avoid hypoglycemia both in the case of insufficient insulin infusion suspension as well as accidental excessive insulin delivery. This approach was used to reduce hypoglycemic risk both during exercise as well as delayed risk during the subsequent night.⁵⁷

We conducted a randomized crossover trial to compare SHAP, DHAP, and the CSII during 24 hours including exercise session (60 minutes at 60% VO_2max) in adults. During the exercise session, hypoglycemia (<60 mg/dL) was common with CSII treatment visits (n = 12) compared to SHAP (n = 3) and DHAP (n = 4). Over the entire study, there was an important reduction in time spent in hypoglycemia with the SHAP compared with CSII, and DHAP halved the time spent below 72 mg/dL compared with the SHAP.⁵⁸ It is noteworthy that exercise was not announced to the algorithm and was preceded by a CHO snack for BG below 90 mg/dL.

Also, our team has compared DHAP against CSII in 15 adults with T1D during an overnight period following an evening exercise session (30 minutes at 60% VO_2max). We have shown that DHAP significantly reduced hypoglycemic risk by up to eightfolds (20-fold reduction overnight) while increasing the percentage of glucose-in-target time by 16% (equivalent to 3.8 hours/day).⁵⁹

Glucagon addition may provide a solution to prevent exercise-induced hypoglycemia. However, the requirement of a pump allowing to separate insulin and glucagon infusion, leading to increased costs and complexity of such devices, in addition to required long-term safety assessment of chronic glucagon use is a significant barrier.^60,61 The balance between benefits and inconvenient (costs and risks) thus remains to be established.

Carbohydrate Supplementation Strategy

The proactive snacking strategy in combination with SHAP would provide a simple and effective means to decrease exercise-induced hypoglycemia in active individuals with T1D.

In 12 patients with T1D, Patel et al⁶⁰ demonstrated that a simple snacking intervention (15 and 30 g CHO for BG >150 and <150 mg/dL, respectively) without announcing exercise could prevent drops in BG around exercise. Without snacking, BG levels decreased by 53±10 mg/dL vs modest increase by 10±13 mg/dL in the snacking strategy (P = .0005) and three patients vs none required rescue CHO treatment (for BG <60 mg/dL). These results support the use of snacking strategy in combination with AP systems to avoid exercise-induced hypoglycemia. Persistent insulin infusion during exercise with snacking also appears to be effective in limiting hyperglycemia at the end of exercise.

Over three days, AP experiments including two sessions (aerobic, resistance, or interval exercise) tested on six subjects, SHAP suggested the consumption of CHO depending on both glucose level and trend.⁶² There were only five hypoglycemia episodes with an average of three hypoglycemia alarms per day per subject. The average CHO consumption was 14±7.8 g per alarm. Around exercise session, CHO amounts are suggested at higher glucose level (80 mg/dL). The authors demonstrated that most hypoglycemia can be predicted, and that low glucose levels can be raised back to the desired levels by consuming an appropriate CHO amount.

It is important to note that snack consumption may be counter-productive given the increased prevalence of the metabolic syndrome in patients with T1D who frequently practice exercise for weight loss or maintenance objectives.^1,3,63

Artificial Pancreas Systems With Combination of Several Strategies

Many studies have attempted to improve the efficiency of the AP systems by combining multiple strategies (glucagon addition, exercise detection, and exercise announcement).

Jacobs et al tested DHAP and exercise announcement at the start of a 45-minute exercise at 60% of their maximum heart rate.⁶⁴ In random order, 21 adults with T1D experienced three sessions: (1) DHAP with exercise announcement; (2) DHAP without exercise announcement; and (3) SAP. Time spent in hypoglycemia during DHAP + announcement was reduced by 2.8% vs unannounced exercise strategy (P = .001) with mild increase in hyperglycemia (primarily postexercise) and no difference vs SAP (P = .16). The authors observed a similar time spent with glucose-in-target between three strategies.⁶⁴

Our group compared the efficacy SHAP vs DHAP during two types of exercise matched for energy expenditure (continuous exercise vs interval exercise).⁶⁵ Exercise was announced 20 minutes prior to its start, and this led to the change of the target glucose level from 95 to 150 mg/dL until the end of the exercise session. Lower doses of glucagon were needed during interval exercise (0.093 ± 0.068 mg) than during continuous exercise (0.126 ± 0.057 mg) (P = .03) due to the lower glucose decrease observed during interval exercise. Overall when combining both exercises, with DHAP, median time spent with glucose-in-target was increased by 28.6% (P = .003) and time in hypoglycemia (BG <70 mg/dL) was decreased by 11% (P = .0001) in comparison to SHAP. The number of hypoglycemia events requiring CHO treatment was also reduced (three in DHAP vs 15 in SHAP), all showing an added benefit of glucagon in the AP systems during exercise.⁶⁵ It can be concluded that the addition of glucagon to the AP systems significantly reduces the risk of hypoglycemia and improves glucose management during both continuous and interval exercise. However, the use of glucagon needs to be further examined under different situations such as proactive snacking and exercise during postabsorptive periods.

Another study was performed in 20 patients with T1D comparing DHAP vs SHAP vs SAP vs usual care for 16 days (four days each arm including 45 minutes at 60% VO_2max).⁶⁶ Both AP systems had an integrated algorithm for exercise detection that received input from heart rate monitor and accelerometer. Mean time in hypoglycemia was the lowest with DHAP during the exercise period: 3.4% ± 4.5% vs 8.3% ± 12.6% for SHAP (P = .009) vs 7.6% ± 8.0% with predictive low glucose suspend (P < .001) vs 4.3% ± 6.8% in the usual care situation where pre-exercise insulin adjustments by the patients were allowed (P = 0.49). Time in hypoglycemia was also the lowest with DHAP during the entire four-day study: 1.3% ± 1.0% vs 2.8% ± 1.7% SHAP (P < .001) vs 2.0% ± 1.5% SAP (P = .04) vs 3.1% ± 3.2% usual care (P = .007). This study showed that the addition of glucagon delivery to a closed-loop system with automated exercise detection reduced hypoglycemia in physically active adults with T1D. It will be interesting to have a similar design with a different exercise type, duration, and intensity.

Conclusion

Studies designed to specifically evaluate the use of AP systems during exercise are limited, but many studies have included periods of structured exercise as part of their experiment protocols that extend to 24 hours or more (summarized in Table 1). AP systems showed better performance than CSII or SAP in many exercise conditions, though complete hypoglycemia prevention is not yet achieved. The major hindrance is related to delays in insulin absorption and interstitial glucose sensing probably making exercise announcement necessary at this stage. Adding biometric variables to detect and classify the type, intensity, and energy expenditure of physical activities can also improve AP systems performance. Yet, the number of wearable devices used should be kept to a minimum and/or integrated to insulin pumps, glucose sensor, or AP controller. Glucagon addition is also a promising avenue, but stable formulations and long-term safety assessment of chronic glucagon use should be first achieved. Ideally, future AP systems, especially with the advances in deep learning artificial intelligence efforts, should be able to predict the effects of specific physical activities (type, intensity, and duration) and changing variables during group sports, in addition to detecting and discriminating exercise-concurrent activities such as meals, sleep, and stress.

Table 1.

Summaries of Main Clinical Trials on Artificial Pancreas Efficiency During Physical Activity in Type 1 Diabetes.

Study	Participants, n	Design, comparators	Artificial pancreas and system specificity	Exercise description	Main outcomes
Haidar et al⁵⁹	Adults, 15	• Crossover, randomized study, two arms • AP systems vs Insulin pump therapy • 16-hour trial	• Addition of glucagon hormone to the AP systems • Exercise unannounced • BG target: 95.4 mg/dL before and after exercise and 150 mg/dL during exercise	• 30 minutes at 60% VO_2max on a stationary bicycle • Postabsorptive state	• No difference in hypoglycemia events or CHO- • requiring related to exercise session • During overall duration (16:00-7:00): • Time spent in hypoglycemia (<72 mg/dL) was lower with DHAP during the day (0.0% [0.0%-3.0%]) vs insulin pump (10.2% [0.0%-13.0%], P = .01), and overnight (0.0% [0.0%-0.0%] vs 0.0% [0.0%-22.5%], P = .02) • More time spent in glycemic target range (70-180 mg/dL) with DHAP during the entire study (70.7% [46.1-88.4]) vs insulin pump (57.3% [25.2-71.8]), P = .003
Jacobs et al⁶⁴	Adults, 21	• Crossover randomized study, three arms • DHAP + exercise announcement vs DHAP + without exercise announcement vs SAP • Three trials, 22-hour sessions	• Addition of glucagon hormone to the AP systems • Exercise announcement • Hormones adjustment during exercise: Insulin was turned off for 30 minutes and then reduced by 50% for 60 minutes. For the DHAP, in addition to the insulin adjustments, the target glucose for glucagon was increased from 95 to 110 mg/dL	• Exercise sessions: 45 minutes at 60% of the maximal HR • Exercise sessions on a treadmill ergometer • Two hours postmeal	• No difference in the time spent in glucose target among all trial arms • Time in hypoglycemia (<70 mg/dL) was less with DHAP + exercise announcement 0.3% vs 3.1% DHAP + unannounced exercise (P = .001) • No difference in time spent in hypoglycemia between DHAP + exercise announcement vs SAP • More time in hyperglycemia (>180 mg/dL) during exercise with DHAP + exercise announcement (32% [IQR 25-39]) vs DHAP + unannounced exercise (25% [19-30]) (P = .001)
Taleb et al⁶⁵	Adults, 17	• Crossover, randomized study, four arms • DHAP vs SHAP during two types of exercise	• Addition of glucagon to the AP systems • Exercise announcement 20 minutes before exercise onset • BG target: 95.4 mg/dL before and after exercise and 150 mg/dL during exercise	• Continuous exercise: 60-minute sessions at 60% VO_2peak • Interval exercise: 40-minute sessions. Two minutes at 85% VO_2peak; two minutes at 50%VO_2peak, two 10-minute periods at 45%VO_2peak (beginning and end of exercise session) • Two hours 30 minutes postmeal	• DHAP is more efficient to prevent the risk of hypoglycemia during both types of exercises • 52.9% had hypoglycemia with SHAP vs 17.6% DHAP (P = .07) during continuous exercise, 46.7% with SHAP during interval training vs 6.25% DHAP (P = .04) • Median time spent with BG <72 mg/dL was lower with DHAP 0% [IQR 0%-0%] vs 22.5% [0-48.3] (P = .006) during continuous exercise and 0% [0%-0%] vs 0% [0%-30.4%] during interval exercise (P = .03) • High glucagon infusion level was needed during continuous exercise compared to interval exercise (P = .03)
Huyett et al⁶⁷	Adolescents, 10	• Crossover study • AP systems vs SAP • Three-day trial	• Announced meals • Unannounced exercise • If the SMBG at the time of the meal is <120 mg/dL the bolus was reduced by 20%. If the SMBG at the time of the meal is >120 mg/dL, a full bolus was given • BG target zone: 80-140 mg/dL (daytime); 90-140 mg/dL (nighttime)	• Two activities per day: participants decided of the type (soccer, tennis, basketball, ultimate frisbee, walking, and bicycling) and intensity of the exercise (low to-moderate intensity) • One to three hours postmeal	• No significant difference in time spent in hypoglycemia with AP systems vs SAP (P = .13) • More time spent in BG target range with AP systems when meals were announced 71% ± 10% vs 0.57% ± 16% SAP (P = .002)
Breton et al⁶⁸	Adolescents, 32	• Crossover, randomized, two arms • AP systems vs SAP • Five-day ski camp • Prolonged physical activity with various intensities and associated metabolic challenges (stress, fear, altitude, and excitement)	• Unannounced exercise	• Five-hour on-snow activity: Morning and afternoon and evening activity after dinner • All activities are in postprandial state	• AP systems improve glycemic control during prolonged intensive exercise • AP group spent more time in glycemic target range (70-180 mg/dL) vs SAP (71.3% ± 17.3% vs 64.7% ± 13.3%), P = .005 • Less time spent in hypoglycemia (<70 mg/dL) with AP (1.6% ± 1.9% [IQR 0-2.3]) vs SAP (3.6% ± 3.2% [0-4.7]), P = .0001 • Subjects had less hypoglycemia with AP systems during skiing sessions regardless of their level: beginners 1.4% vs SAP 2.3% and advanced 1.3% vs SAP 2.3%
DeBoer et al⁶⁹	Adolescents, 18	• Crossover, randomized study, two arms • AP systems + HRM vs AP − HRM • One-day trial	• Exercise detected with HRM	• 15 minutes at submaximal exercise bouts on a cycle ergometer • Five minutes of rest between each bout • Postabsorptive state	• No difference in time spent in normoglycemia (70-180 mg/dL) between AP + HRM and AP − HRM (P = .366) overall but more time in normoglycemia during exercise and rest periods with AP + HRM (96% vs 87% AP − HRM, P = .075) • AP + HRM reduced the time spent in hypoglycemia during exercise and rest (0.5% vs 7.4% AP − HRM, P = .028).
De Bock et al⁷⁰	Adults and adolescents, 8	• AP systems vs AP systems + overreading sensor vs AP systems with overreading sensor + exercise • Four-day trial • One exercise session on day 3	• AP systems with insulin delivery limit • Announced meals • Unannounced exercise • Overcalibration of the glucose sensor by 20% on day 2 and maintained on day 3 with exercise • Meal restriction on day 3	• Aerobic exercise, 45 minutes, at 55% of VO_2max on a stationary bicycle • Four hours postmeal	• Interruption of insulin delivery alone was not enough to prevent hypoglycemia • Hypoglycemic events occurred on average 115 ± 50 minutes from last bolus • Insulin delivery was stopped 110 ± 53 minutes before hypoglycemia
Castle et al ⁶⁶	Adults, 25	• Crossover, randomized study, four arms • SHAP vs DHAP vs predictive low glucose suspend vs usual care • Each arm included three exercise sessions	• Addition of glucagon to the AP systems • Automatic exercise detection algorithm with exogenous inputs: (HRM and accelerometer) when METs >4 for five consecutive minutes followed by user confirmation • Hormones adjustment during exercise: insulin was turned off for 30 minutes and then reduced by 50% for 60 minutes. For the DHAP, in addition to the insulin adjustments, the BG target for glucagon was increased from 95 to 120 mg/dL	• Aerobic exercise, 45 minutes, at 60% of VO_2max • Two hours postmeal	• Less time spent in hypoglycemia during exercise with DHAP 3.4% ± 1.4% vs 8.3% ± 12.6% SHAP (P < .009), vs 7.6% ± 8.0% predictive low glucose suspend (P < .001), vs 4.3% ± 6.8% usual care (P = .49) • Less time spent in hypoglycemia during the entire study period with DHAP (1.3% vs 2.8% SHAP, P < .001) • No significative difference in sensitivity and specificity of exercise detection between SHAP and DHAP
Turksoy et al⁵⁴	Adults, 10	• Crossover study • Challenged AP systems during three-day trial; 18 visits with various exercise types and intensities • Two exercise sessions per day	• Meal detection and CHO recommendation algorithm • Automatic exercise detection with exogenous inputs: accelerometer reporting energy expenditure and galvanic skin response • Hypoglycemia early alarm system that predicts glucose concentrations and notifies patients to act when necessary to prevent hypoglycemia • BG target: 120 mg/dL and remains unchanged during exercise	• Day 1 and 3: • Aerobic, MCT: warm-up, five minutes; 20 minutes at 60%-80% of HR reserve; five minutes cool down. • RT: four exercises, three sets, ten repetitions • Day 2: RT and HIIT: five minutes warm up; three minutes at 60%-70% HR; four minutes at 80%-90% HR; all repeated three times	• The average BG levels decreased consistently during exercise and recovery but were different depending on the type of exercise (P = .0097) with a significant difference between MCT and RT (P = .0075) and no significant difference between MCT and HIIT or RT and HIIT (P = .521 and P = .325, respectively) • AP systems maintained BG values within the target range for 69.92% ± 12.79% of the time all types of exercises included
Pinsker et al⁷¹	Adults, 15	• Controlled study • AP systems vs SAP • 48-hour trial preceded by a one-week run-in phase	• AP systems modified meal bolus based on the CGM value (BG <110 mg/dL meal bolus reduced by 20%; between 110 and 150 mg/dL a full meal bolus was given. BG >150 mg/dL a full meal bolus plus two units correction) • Exercise announcement • BG target is unchanged during exercise	• One-hour outdoor exercise session (brisk walk) • Three hours postmeal	• AP systems are efficient to maintain BG values within the desired target range (70-180 mg/dL) • Glucose values were in the desired target range 88.0% ± 8.0% of the time during AP sessions vs 74.6% ± 9.0% SAP, P = .001 • Time spent in hypoglycemia (<70 mg/dL) decreased from 7.8% ± 6.0% with SAP to 1.5% ± 1.9% with AP systems (P < .001)
Wood et al⁷²	Children, 111	• AP systems • Six days/five nights trial for three months preceded by a two-week run-in phase	• Insulin suspension if predicted risk of hypoglycemia with low limit set at 65 mg/dL • Exercise unannounced	• Aerobic activities: walking, biking, and playing video games • Between 6:30 and 8:00 pm	• Hypoglycemia was avoided 79.7% of the time when the suspend before low system was activated without rebound hyperglycemia • On average, insulin was suspended during 87.5 ± 32.7 minutes • No severe hypoglycemic events occurred
Quirós et al⁷³	Adults, 5	• Crossover study, two arms • AP systems with aerobic exercise vs AP systems with anaerobic exercise	• Exercise announced • BG target: 100 mg/dL before and after exercise and 150 mg/dL during exercise	• Postmeal • Aerobic exercise: Three series of 15 minutes on a stationary bicycle at 60% of HR_max. Five minutes rest between each series • Anaerobic exercise: Four exercises. Five series of eight repetitions at 70% of maximal capacity	• Glucose values were within the desired target range (70-180 mg/dL), 89.8% ± 18.6% of the time during aerobic exercise vs 75.9% ± 27.6% during resistance exercise • Time spent in hypoglycemia (<70 mg/dL) was 2.5% ± 6.3% during aerobic exercise and 1% ± 3.6% during resistance exercise
Forlenza et al⁷⁴	Adults, 12	• Crossover study, two arms • AP systems with raised set points vs AP systems with reduced basal insulin • 54-hour trial	• Variable glucose set points and basal rate: • From 130 to 150 mg/dL, 90 minutes before exercise with basal rate unchanged • 130 mg/dL, 90 minutes before exercise with 50% of basal rate	• Moderate intensity exercise >30 minutes, afternoon • Activities: soccer, basketball, and treadmill	• Good glycemic control was maintained during moderate intensity exercise with changing glucose targets or modified basal • No significant difference in the time spent within the 70-180 mg/dL target range between raised set points 88.9% ± 17.6% and reduced basal insulin 89.1% ± 11.3% nor in the time spent in hypoglycemia (<70 mg/dL), 1.4% ± 2.7% with raised set point vs 1.3% ± 3.0% with reduced basal insulin during the 12-hour period after exercise
Ramkissoon et al⁷⁵	Adults, 10	• Crossover study, three arms • In silico, regular AP systems + announced exercise vs regular AP systems + unannounced exercise vs enhanced AP systems with automatic exercise detection • 15-day trial scenario	• AP systems with automatic exercise detection: • Automatic exercise detection with exogenous input: accelerometer reporting energy expenditure • Exercise detection triggers three actions: reducing basal insulin to 0; reducing meal insulin bolus by 30%; suggesting a specific amount of CHO to the patient	• Eight aerobic exercise sessions (60% VO_2max) on alternative days • Exercise tested for pre- and postbreakfast, pre- and postlunch, in the afternoon and evening	• AP systems + automatic detection managed to prevent severe hypoglycemia for every intensity of exercise tested • Less time spent in hypoglycemia with AP systems + automatic detection 0.1% [0.0-0.1] vs 1.9% [1.5-2.8] with AP systems + unannounced exercise (P = .002), and vs 0.4% [0.2-0.6] with AP systems + announced exercise (P = .006) • No difference in time spent in hyperglycemia between AP systems + automatic detection vs AP systems + announced (P = .432) • More time spent in hyperglycemia with AP systems + automatic detection 6.4% [6.0-8.7] vs 4.2% [2.0-4.6] with AP systems + unannounced exercise (P = .002)
Paldus et al⁷⁶	Adults, 11	• Crossover, randomized study, two arms • Standard AP systems vs enhanced AP systems	• AP systems challenged with • two missed meal boluses, two high glycemic index, and two high fat meals; exercise	• 40-minute moderate intensity exercise	• Overall, patients were more satisfied with the enhanced AP systems • Less alerts occurred with the enhanced AP systems 3.9 ± 2.8 vs 8.6 ± 5.8 with standard AP (P = .01) • No differences regarding glucose control were observed in the two arms

Abbreviations: AP, artificial pancreas; BG, blood glucose; CHO, carbohydrate; DHAP, dual-hormone artificial pancreas; HR, heart rate; HRM, heart rate monitor; HRR, heart rate reserve; MCT, moderate continuous training; MET, metabolic equivalent; RT, resistance training; SAP, sensor augmented pump therapy; SHAP, single-hormone artificial pancreas; SMBG, self-monitoring blood glucose; VO_2max, maximal oxygen uptake; IQR, interquartile range.

Footnotes

Acknowledgements

NT is the recipient of scholarship of CIHR (Canadian Institutes of Health Research) and FRSQ (Fonds de Recherche Santé Québec) scholarships. RRL is holding the J-A DeSève diabetes research chair and a Diabetes Canada program grant (DI-5-17-5282-RR) that both supported this work.

Authors’ Contributions

ST, NT, and RRL conceived the literature review design and content. ST drafted the manuscript which was critically reviewed by NT, JM, EN, MR, and RRL. All authors revised and agreed upon the final submitted manuscript.

Declaration of Conflicting Interests

The author(s) declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article: ST NT, JM, EN, and MR have no conflicts of interest to declare. RRL has the following to declare: Research grants from Astra-Zeneca, E Lilly, Merck, NIH, Novo-Nordisk, Sanofi-Aventis; Consulting/advisory panel with Abbott, Amgen, Astra-Zeneca, Boehringer, Carlina Technology, Eli Lilly, Janssen, Medtronic, Merck, Neomed, Novo-Nordisk, Roche, Sanofi-Aventis; Honoraria for conferences received from Abbott, Astra-Zeneca, Boehringer, E Lilly, Janssen, Medtronic, Merck, Novo-Nordisk, Sanofi-Aventis; consumable gift (in Kind) from Abbott, Animas, Medtronic, Roche; unrestricted grants for clinical and educational activities from Eli Lilly, Lifescan, Medtronic, Merck, Novo Nordisk, Sanofi; patent for T2DM risk biomarkers, catheter life & artificial pancreas; Purchase fees related to artificial pancreas from Eli Lilly.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by Diabetes Canada program grant (DI-5-17-5282-RR). ST is supported by a fellowship grant from SFD (Société Francophone du Diabète) and FRM (Fondation de Recherche Médicale).

ORCID iD

Sémah Tagougui

References

Riddell

Gallen

Smart

et al. Exercise management in type 1 diabetes: a consensus statement. Lancet Diabetes Endocrinol. 2017;5(5):377-390.

Chimen

Kennedy

Nirantharakumar

Pang

Andrews

Narendran

What are the health benefits of physical activity in type 1 diabetes mellitus? A literature review. Diabetologia. 2012;55(3):542-551.

Tagougui

Taleb

Rabasa-Lhoret

The benefits and limits of technological advances in glucose management around physical activity in patients type 1 diabetes. Front Endocrinol. 2018;9:818.

Rabasa-Lhoret

Bourque

Ducros

Chiasson

JL.

Guidelines for premeal insulin dose reduction for postprandial exercise of different intensities and durations in type 1 diabetic subjects treated intensively with a basal-bolus insulin regimen (ultralente-lispro). Diabetes Care. 2001;24(4):625-630.

Zaharieva

McGaugh

Pooni

Vienneau

Riddell

MC.

Improved open-loop glucose control with basal insulin reduction 90 minutes before aerobic exercise in patients with type 1 diabetes on continuous subcutaneous insulin infusion. Diabetes Care. 2019; 42(5):824-831.

Brazeau

A-S

Rabasa-Lhoret

Strychar

Mircescu

Barriers to physical activity among patients with type 1 diabetes. Diabetes Care. 2008;31(11):2108-2109.

Brazeau

Mircescu

Desjardins

et al. The barriers to physical activity in type 1 diabetes (BAPAD-1) scale: predictive validity and reliability. Diabetes Metab. 2012;38(2):164-170.

Tansey

Tsalikian

Beck

et al. The effects of aerobic exercise on glucose and counterregulatory hormone concentrations in children with type 1 diabetes. Diabetes Care. 2006;29(1):20-25.

Galassetti

Riddell

MC.

Exercise and type 1 diabetes (T1DM). Compr Physiol. 2013;3(3):1309-1336.

10.

Yardley

Colberg

SR.

Update on management of type 1 diabetes and type 2 diabetes in athletes. Curr Sports Med Rep. 2017;16(1):38-44.

11.

Yardley

Brockman

Bracken

RM.

Could age, sex and physical fitness affect blood glucose responses to exercise in type 1 diabetes?

Front Endocrinol. 2018;9:674.

12.

Facchinetti

Del Favero

Sparacino

Cobelli

Model of glucose sensor error components: identification and assessment for new Dexcom G4 generation devices. Med Biol Eng Comput. 2015;53(12):1259-1269.

13.

Taplin

Cobry

Messer

McFann

Chase

Fiallo-Scharer

Preventing post-exercise nocturnal hypoglycemia in children with type 1 diabetes. J Pediatr. 2010;157(5):784-788.e1. doi:10.1016/j.jpeds.2010.06.004

14.

Taleb

Emami

Suppere

et al. Comparison of two continuous glucose monitoring systems, Dexcom G4 Platinum and Medtronic paradigm Veo Enlite System, at rest and during exercise. Diabetes Technol Ther. 2016;18(9):561-567.

15.

Bally

Zueger

Pasi

Carlos

Paganini

Stettler

Accuracy of continuous glucose monitoring during differing exercise conditions. Diabetes Res Clin Pract. 2016;112:1-5.

16.

Iscoe

Campbell

Jamnik

Perkins

Riddell

MC.

Efficacy of continuous real-time blood glucose monitoring during and after prolonged high-intensity cycling exercise: spinning with a continuous glucose monitoring system. Diabetes Technol Ther. 2006;8(6):627-635.

17.

Moser

Mader

Tschakert

et al. Accuracy of continuous glucose monitoring (CGM) during continuous and high-intensity interval exercise in patients with type 1 diabetes mellitus. Nutrients. 2016;8(8): 489.

18.

Yardley

Sigal

Kenny

Riddell

Lovblom

Perkins

BA.

Point accuracy of interstitial continuous glucose monitoring during exercise in type 1 diabetes. Diabetes Technol Ther. 2013;15(1):46-49.

19.

Younk

Mikeladze

Tate

Davis

SN.

Exercise-related hypoglycemia in diabetes mellitus. Expert Rev Endocrinol Metab. 2011;6(1):93-108.

20.

Gerich

Langlois

Noacco

Karam

Forsham

PH.

Lack of glucagon response to hypoglycemia in diabetes: evidence for an intrinsic pancreatic alpha cell defect. Science. 1973;182(4108):171-173.

21.

Colberg

Laan

Dassau

Kerr

Physical activity and type 1 diabetes: time for a rewire?

J Diabetes Sci Technol. 2015;9(3):609-618.

22.

Fahey

Paramalingam

Davey

Davis

Jones

Fournier

PA.

The effect of a short sprint on postexercise whole-body glucose production and utilization rates in individuals with type 1 diabetes mellitus. J Clin Endocrinol Metab. 2012;97(11):4193-4200.

23.

Yardley

Sigal

Riddell

Perkins

Kenny

GP.

Performing resistance exercise before versus after aerobic exercise influences growth hormone secretion in type 1 diabetes. Appl Physiol Nutr Metab. 2014;39(2):262-265.

24.

Yardley

Kenny

Perkins

et al. Resistance versus aerobic exercise: acute effects on glycemia in type 1 diabetes. Diabetes Care. 2013;36(3):537-542. doi:10.2337/dc12–0963

25.

Sigal

Purdon

Fisher

Halter

Vranic

Marliss

EB.

Hyperinsulinemia prevents prolonged hyperglycemia after intense exercise in insulin-dependent diabetic subjects. J Clin Endocrinol Metab. 1994;79(4):1049-1057.

26.

Purdon

Brousson

Nyveen

et al. The roles of insulin and catecholamines in the glucoregulatory response during intense exercise and early recovery in insulin-dependent diabetic and control subjects. J Clin Endocrinol Metab. 1993;76(3):566-573.

27.

Aronson

Brown

Riddell

MC.

Optimal insulin correction factor in post-high-intensity exercise hyperglycemia in adults with type 1 diabetes: the FIT study. Diabetes Care. 2019;42(1):10-16.

28.

Sandoval

Guy

DLA

Richardson

Ertl

Davis

SN.

Effects of low and moderate antecedent exercise on counterregulatory responses to subsequent hypoglycemia in type 1 diabetes. Diabetes. 2004;53(7):1798-1806.

29.

American Diabetes Association. Physical activity/exercise and diabetes. Diabetes Care. 2004;27(Suppl 1):S58-S62.

30.

Kemmer

FW.

Prevention of hypoglycemia during exercise in type I diabetes. Diabetes Care. 1992;15(11):1732-1735.

31.

Grimm

Ybarra

Berné

Muchnick

Golay

A new table for prevention of hypoglycaemia during physical activity in type 1 diabetic patients. Diabetes Metab. 2004;30(5):465-470.

32.

Riddell

Bar-Or

Ayub

Calvert

Heigenhauser

GJ.

Glucose ingestion matched with total carbohydrate utilization attenuates hypoglycemia during exercise in adolescents with IDDM. Int J Sport Nutr. 1999;9(1):24-34.

33.

Francescato

Geat

Accardo

Blokar

Cattin

Noacco

Exercise and glycemic imbalances: a situation-specific estimate of glucose supplement. Med Sci Sports Exerc. 2011;43(1):2-11.

34.

Mauvais-Jarvis

Sobngwi

Porcher

et al. Glucose response to intense aerobic exercise in type 1 diabetes: maintenance of near euglycemia despite a drastic decrease in insulin dose. Diabetes Care. 2003;26(4):1316-1317.

35.

Rabasa-Lhoret

Garon

Langelier

Poisson

Chiasson

JL.

Effects of meal carbohydrate content on insulin requirements in type 1 diabetic patients treated intensively with the basal-bolus (ultralente-regular) insulin regimen. Diabetes Care. 1999;22(5):667-673.

36.

Roy-Fleming

Taleb

Messier

et al. Timing of insulin basal rate reduction to reduce hypoglycemia during late post-prandial exercise in adults with type 1 diabetes using insulin pump therapy: a randomized crossover trial [published online ahead of print August 27, 2018]. Diabetes Metab. doi:10.1109/TBME.2017.2652062

37.

Bakhtiani

Zhao

El Youssef

Castle

Ward

WK.

A review of artificial pancreas technologies with an emphasis on bi-hormonal therapy. Diabetes Obes Metab. 2013;15(12):1065-1070.

38.

Bertachi

Ramkissoon

Bondia

Vehí

Automated blood glucose control in type 1 diabetes: a review of progress and challenges. Endocrinol Diabetes Nutr. 2018;65(3):172-181.

39.

Weisman

Bai

J-W

Cardinez

Kramer

Perkins

BA.

Effect of artificial pancreas systems on glycaemic control in patients with type 1 diabetes: a systematic review and meta-analysis of outpatient randomised controlled trials. Lancet Diabetes Endocrinol. 2017;5(7):501-512.

40.

Riddell

Zaharieva

Yavelberg

Cinar

Jamnik

VK.

Exercise and the development of the artificial pancreas: one of the more difficult series of hurdles. J Diabetes Sci Technol. 2015;9(6):1217-1226.

41.

Zaharieva

Yavelberg

Jamnik

Cinar

Turksoy

Riddell

MC.

The effects of basal insulin suspension at the start of exercise on blood glucose levels during continuous versus circuit-based exercise in individuals with type 1 diabetes on continuous subcutaneous insulin infusion. Diabetes Technol Ther. 2017;19(6):370-378.

42.

Gingras

Bonato

Messier

et al. Impact of macronutrient content of meals on postprandial glucose control in the context of closed-loop insulin delivery: a randomized cross-over study. Diabetes Obes Metab. 2018;20(11):2695-2699.

43.

Breton

MD.

Handling exercise during closed loop control. Diabetes Technol Ther. 2017;19(6):328-330.

44.

Konstantatos

Kavnoudias

Stegeman

et al. A randomized, double-blind, placebo-controlled study of preemptive oral oxycodone with morphine patient-controlled anesthesia for postoperative pain management in patients undergoing uterine artery embolization for symptomatic uterine fibroids. Cardiovasc Intervent Radiol. 2014;37(5):1191-1197.

45.

Cobelli

Renard

Kovatchev

Artificial pancreas: past, present, future. Diabetes. 2011;60:2672-2682.

46.

Regittnig

Ellmerer

Fauler

et al. Assessment of transcapillary glucose exchange in human skeletal muscle and adipose tissue. Am J Physiol Endocrinol Metab. 2003;285(2):E241-E251.

47.

Rebrin

Sheppard

Steil

GM.

Use of subcutaneous interstitial fluid glucose to estimate blood glucose: revisiting delay and sensor offset. J Diabetes Sci Technol. 2010;4(5):1087-1098.

48.

Schiavon

Dalla Man

Kudva

Basu

Cobelli

In silico optimization of basal insulin infusion rate during exercise: implication for artificial pancreas. J Diabetes Sci Technol. 2013;7(6):1461-1469.

49.

Turksoy

Monforti

Park

Griffith

Quinn

Cinar

Use of wearable sensors and biometric variables in an artificial pancreas system. Sensors. 2017;17(3) : 532

50.

Turksoy

Paulino

TML

Zaharieva

et al. Classification of physical activity: information to artificial pancreas control systems in real time. J Diabetes Sci Technol. 2015;9(6):1200-1207.

51.

Elleri

Allen

Nodale

et al. Automated overnight closed-loop glucose control in young children with type 1 diabetes. Diabetes Technol Ther. 2011;13(4):419-424.

52.

Breton

Farret

Bruttomesso

et al. Fully integrated artificial pancreas in type 1 diabetes: modular closed-loop glucose control maintains near normoglycemia. Diabetes. 2012;61(9):2230-2237.

53.

Biagi

Bertachi

Quirós

et al. Accuracy of continuous glucose monitoring before, during, and after aerobic and anaerobic exercise in patients with type 1 diabetes mellitus. Biosensors. 2018;8(1): 22

54.

Turksoy

Hajizadeh

Hobbs

et al. Multivariable artificial pancreas for various exercise types and intensities. Diabetes Technol Ther. 2018;20(10):662-671.

55.

Jacobs

Resalat

El Youssef

et al. Incorporating an exercise detection, grading, and hormone dosing algorithm into the artificial pancreas using accelerometry and heart rate. J Diabetes Sci Technol. 2015;9(6):1175-1184.

56.

Franc

Daoudi

Pochat

et al. Insulin-based strategies to prevent hypoglycaemia during and after exercise in adult patients with type 1 diabetes on pump therapy: the DIABRASPORT randomized study. Diabetes Obes Metab. 2015;17(12):1150-1157.

57.

Castle

Engle

El Youssef

et al. Novel use of glucagon in a closed-loop system for prevention of hypoglycemia in type 1 diabetes. Diabetes Care. 2010;33(6):1282-1287.

58.

Haidar

Legault

Matteau-Pelletier

et al. Outpatient overnight glucose control with dual-hormone artificial pancreas, single-hormone artificial pancreas, or conventional insulin pump therapy in children and adolescents with type 1 diabetes: an open-label, randomised controlled trial. Lancet Diabetes Endocrinol. 2015;3(8):595-604.

59.

Haidar

Legault

Dallaire

et al. Glucose-responsive insulin and glucagon delivery (dual-hormone artificial pancreas) in adults with type 1 diabetes: a randomized crossover controlled trial. CMAJ. 2013;185(4):297-305.

60.

Patel

Van Name

Cengiz

et al. Mitigating reductions in glucose during exercise on closed-loop insulin delivery: the ex-snacks study. Diabetes Technol Ther. 2016;18(12):794-799.

61.

El-Khatib

Russell

Magyar

et al. Autonomous and continuous adaptation of a bihormonal bionic pancreas in adults and adolescents with type 1 diabetes. J Clin Endocrinol Metab. 2014;99(5):1701-1711.

62.

Turksoy

Kilkus

Hajizadeh

et al. Hypoglycemia detection and carbohydrate suggestion in an artificial pancreas. J Diabetes Sci Technol. 2016;10(6):1236-1244.

63.

Leroux

Brazeau

A-S

Gingras

Desjardins

Strychar

Rabasa-Lhoret

Lifestyle and cardiometabolic risk in adults with type 1 diabetes: a review. Can J Diabetes. 2014;38(1):62-69.

64.

Jacobs

El Youssef

Reddy

et al. Randomized trial of a dual-hormone artificial pancreas with dosing adjustment during exercise compared with no adjustment and sensor-augmented pump therapy. Diabetes Obes Metab. 2016;18(11):1110-1119.

65.

Taleb

Emami

Suppere

et al. Efficacy of single-hormone and dual-hormone artificial pancreas during continuous and interval exercise in adult patients with type 1 diabetes: randomised controlled crossover trial. Diabetologia. 2016;59(12):2561-2571.

66.

Castle

El Youssef

Wilson

et al. Randomized outpatient trial of single-and dual-hormone closed-loop systems that adapt to exercise using wearable sensors. Diabetes Care. 2018;41(7):1471-1477.

67.

Huyett

Forlenza

et al. Outpatient closed-loop control with unannounced moderate exercise in adolescents using zone model predictive control. Diabetes Technol Ther. 2017;19(6):331-339.

68.

Breton

Cherñavvsky

Forlenza

et al. Closed-loop control during intense prolonged outdoor exercise in adolescents with type 1 diabetes: the artificial pancreas ski study. Diabetes Care. 2017;40(12):1644-1650.

69.

DeBoer

Cherñavvsky

Topchyan

Kovatchev

Francis

Breton

MD.

Heart rate informed artificial pancreas system enhances glycemic control during exercise in adolescents with T1D. Pediatr Diabetes. 2017;18(7):540-546.

70.

de Bock

Dart

Roy

et al. Exploration of the performance of a hybrid closed loop insulin delivery algorithm that includes insulin delivery limits designed to protect against hypoglycemia. J Diabetes Sci Technol. 2017;11(1):68-73.

71.

Pinsker

Laguna Sanz

Lee

et al. Evaluation of an artificial pancreas with enhanced model predictive control and a glucose prediction trust index with unannounced exercise. Diabetes Technol Ther. 2018;20(7):455-464.

72.

Wood

Shulman

Forlenza

et al. In-clinic evaluation of the MiniMed 670G system “Suspend Before Low” feature in children with type 1 diabetes. Diabetes Technol Ther. 2018;20(11):731-737.

73.

Quirós

Bertachi

Giménez

et al. Blood glucose monitoring during aerobic and anaerobic physical exercise using a new artificial pancreas system. Endocrinol Diabetes Nutr. 2018;65(6):342-347.

74.

Forlenza

Buckingham

Christiansen

et al. Performance of omnipod personalized model predictive control algorithm with moderate intensity exercise in adults with type 1 diabetes. Diabetes Technol Ther. 2019;21(5):265-272. doi: 10.1089/dia.2019.0017

75.

Ramkissoon

Bertachi

Beneyto

Bondia

Vehi

. Detection and control of unannounced exercise in the artificial pancreas without additional physiological signals [published online ahead of print February 11, 2019]. IEEE J Biomed Health Inform. doi: 10.1109/JBHI.2019.2898558

76.

Paldus

Lee

Jones

et al. Glucose control using a standard versus an enhanced hybrid closed loop system: a randomized crossover study. Diabetes Technol Ther. 2019;21(1):56-58.