Exercise Strategies for Hypoglycemia Prevention in Individuals With Type 1 Diabetes

By definition, aerobic exercise (e.g., swimming, jogging, cycling, or walking) involves the continuous and repeated movements of large muscle groups for at least 10 minutes at a time (24). The initial stages of this type of exercise are fueled mostly by muscle glycogen, after which glucose and nonesterified fatty acids become the main source of fuel. In individuals without diabetes, insulin release from pancreatic β-cells decreases, and glucagon secretion increases during moderate-intensity aerobic exercise. These changes ensure that fatty acids (stored in adipocytes) and glucose (stored mainly in the liver as glycogen) are released from storage to fuel the exercising muscles. As a result, blood glucose levels can remain relatively stable during exercise for several hours without food intake.

For individuals with type 1 diabetes, the lack of β-cell function leads to the requirement for exogenous insulin (introduced into the body by injection or infusion). Circulating levels of insulin consequently cannot be regulated endogenously and depend on the quantity and timing of insulin taken by the individual before exercise. Insulin levels are often higher than they would be in the absence of diabetes, which has the result of limiting glucose production by the liver while stimulating glucose uptake by muscle, adipose, and liver cells for storage. As a result, blood glucose levels often decrease dramatically during physical activity for individuals with type 1 diabetes unless carbohydrates are consumed before, during, and after exercise.

As exercise intensity increases, the relative roles of insulin and glucagon decrease (25). Glucose breakdown becomes rapid and exceeds the capacity of the oxidative systems. This results in the breakdown of glucose to lactate without the use of oxygen in the muscles (anaerobic glycolysis). Circulating levels of catecholamines (epinephrine and nonepinephrine) increase substantially and exert an overriding influence on glucose release by the liver (26). This can lead to glucose production exceeding demand, resulting in an increase in blood glucose concentration (26,27).

High-intensity exercise is also associated with increases in growth hormone and lactate, both of which can serve to spare circulating blood glucose. Growth hormone promotes lipolysis, potentially decreasing the body’s reliance on circulating glucose, and lactate is a gluconeogenic precursor that can be used by the liver to create new glucose. Intuitively, it thus makes sense that any situation increasing the circulation of these four compounds in the blood would be associated with either increases or attenuated declines in blood glucose levels.

Short bursts of high-intensity exercise, in the form of short sprints, intermittent, high-intensity (>85% of peak aerobic capacity [VO₂peak]) intervals, and resistance exercise (weight lifting), have consequently become a topic of interest to those in the diabetes research community because it has been suggested that the inclusion of these types of activities in exercise sessions involving moderate aerobic exercise may provide a level of protection against hypoglycemia apart from that already demonstrated by increases in carbohydrate consumption and decreases in insulin dosage. Table 1 lists the main hormonal and blood glucose responses to aerobic and sprint exercise in people with and without type 1 diabetes.

Similar to individuals without diabetes, short sprints and intermittent, high-intensity exercise in people with type 1 diabetes are associated with increases in circulating epinephrine, norepinephrine, growth hormone, and lactate when compared to aerobic exercise performed on its own (15,16,18–21,28). A series of studies have used this knowledge to explore the possibility that including a simple all-out 10-second sprint either before (16) or after (15,17) a 20- to 30-minute bout of moderate aerobic exercise (cycling at 40% of VO₂peak) could protect against exercise-induced hypoglycemia in individuals with type 1 diabetes. Bussau et al. (16) found that performing a 10-second sprint before 20 minutes of moderate-intensity (40% of VO₂peak) cycling did not protect against declines in blood glucose during exercise in seven individuals with type 1 diabetes but did prevent further declines in blood glucose in the 45 minutes after exercise compared to an exercise session during which no sprint was performed. When the 10-second sprint was performed at the completion of a similar 20-minute, moderate-intensity aerobic exercise session, there was no significant change in blood glucose levels for the subsequent 2 hours after exercise compared to a decline of 65 ± 22 mg/dL measured after 20 minutes of exercise during which a sprint was not included (15). A follow-up study of the 10-second sprint (performed in isolation) found that the increases in blood glucose concentration arising from the sprint resulted from a decrease in the rate of glucose uptake rather than a disproportionate increase in glucose appearance (19). The authors concluded that a likely cause of this phenomenon was a build-up of intramuscular glucose-6-phosphate (associated with the rapid glycogen breakdown that would be found during a sprint), which would result in inhibition of the enzyme hexokinase and subsequently a decrease in muscle glucose uptake (19).

A handful of studies in the past decade have examined the use of repeated bouts of intermittent, high-intensity exercise (4- to 15-second intervals at >85% VO₂peak performed every 2–5 minutes) for preventing hypoglycemia both during and after exercise in individuals with type 1 diabetes (20,21,28–31). In attempting to mimic the demands of field sports, Guelfi et al. (20) asked seven young, healthy individuals with type 1 diabetes to perform two separate exercise sessions. One session involved moderate-intensity (40% of VO₂peak) cycling for 30 minutes, whereas the other session involved an identical intensity and duration of cycling with 4-second sprints performed every 2 minutes throughout exercise (20). Despite the higher energy demands of the exercise session that included sprints, declines in blood glucose during exercise were lower when the sprints were included (–52 ± 14 vs. 79 ± 22 mg/dL, P = 0.006) (20). In addition, whereas blood glucose levels continued to decline for 60 minutes post-exercise after the moderate-intensity session, further declines in blood glucose did not occur after the completion of the session in which intermittent sprints were included (20). A follow-up study using the same protocol (21) further elucidated that the slower decline in blood glucose levels during the intermittent sprint session could be attributed to an increased rate of glucose appearance early in the exercise session compared to the moderate exercise session, whereas glucose disappearance in the intermittent sprint session was attenuated during and after exercise compared to the moderate-intensity session. Both studies found an increase in epinephrine and norepinephrine by the end of exercise (20,21), which may have contributed to higher rates of glucose appearance through an increase in hepatic glucose production (32) and a decrease in glucose disposal.

A similar study of intermittent, high-intensity activity performed by Maran et al. (28) showed a statistically nonsignificant trend toward smaller declines in blood glucose when eight participants with type 1 diabetes performed 30 minutes of moderate (40% VO₂peak) cycling with 5-second intervals at 85% VO₂peak performed every 2 minutes compared to when the same participants performed continuous moderate cycling. In contrast, Dubé et al. (31) found that changes in blood glucose during exercise were similar when 11 participants with type 1 diabetes performed 60 minutes of moderate (50% VO₂peak) cycling and when the same participants performed 60 minutes of moderate cycling with 10-second sprints included every 2 minutes. It is important to note, however, that twice as many participants required glucose infusion both during (7/11 vs. 4/11) and after exercise (7/11 vs. 3/11) in the moderate exercise session compared to the intermittent, high-intensity session (31). Similarly, when high-intensity intervals (15-second maximal sprints) were spaced farther apart (5 minutes) during 45 minutes of cycling at 50% of the subjects’ maximum work rate in a study by Iscoe and Riddell (29), changes in blood glucose were very similar to those found in a continuous moderate cycling session. Once again, however, more carbohydrate intake was required to prevent hypoglycemia during the moderate exercise session compared to the session in which high-intensity intervals were included (29).

Only one study to date has examined the effects of intermittent, high-intensity running on blood glucose levels in individuals with type 1 diabetes. Campbell et al. (30) argued that intermittent, high-intensity cycling protocols, which rely mainly on concentric muscle action, fail to adequately replicate the physiological demands of repeated changes in speed and direction involved in field and ice sports, which involve a significant amount of eccentric muscle work. Consequently, they decided to compare the effects of the Loughborough Intermittent Shuttle Test (33) performed for 45 minutes to the effects of a 45-minute moderate- to high-intensity (∼77% VO₂peak) treadmill run in nine individuals with type 1 diabetes. The shuttle test consisted of the following sequence: 3 × 20 meter shuttle walking and 1 × 20 meter sprinting with a 4-second rest between each, followed by 3 × 20 meters at a speed equivalent to 55% VO₂peak and 3 × 20 meters at a speed corresponding to 95% VO₂peak (30). The change in blood glucose during exercise was approximately four times greater during the continuous exercise session compared to the shuttle test. During the hour post-exercise, declines in blood glucose after the shuttle test were also significantly smaller than those found during the aerobic running session (–20 ± 25 vs. –95 ± 7.2 mg/dL, P = 0.037) (30). The authors suggested that the tenfold higher lactate levels measured at the end of the shuttle test likely contributed substantially to the attenuated rates of blood glucose decline during the 60 minutes post-exercise by increasing gluconeogenesis over this period.

These studies indicate that the inclusion of short sprints and intermittent, high-intensity exercise (lasting 5–15 seconds) into aerobic exercise sessions may be an effective way to counter declines in blood glucose during and shortly after exercise. In addition, although antecedent hypoglycemia may increase the risk of subsequent exercise-induced hypoglycemia (34–36), there is evidence to indicate that the blood glucose–raising effect of a short sprint is not impaired by prior hypoglycemia (18) and may therefore still be effective at maintaining blood glucose levels. As opposed to moderate-intensity exercise after antecedent hypoglycemia, which blunts cortisol, epinephrine, and glucagon responses (34,36) and thereby increases the risk of exercise-induced hypoglycemia, one study has shown that these responses are intact when a 10-second, all-out sprint is performed after a similar low blood glucose event (18). It should be noted, however, that antecedent exercise of sufficient intensity and duration to deplete glycogen stores may decrease the amount of glucose available and therefore diminish the protective effect of a sprint. The magnitude of this dampening is currently unknown, and there are no published studies on this topic to date.

Resistance exercise, which involves lifting weights or working against a resistive force, also results in increases in epinephrine (37,38), norepinephrine (37,38), lactate (38), and growth hormone (39), which could help in stabilizing blood glucose levels in a manner similar to other forms of anaerobic exercise (23,40). The type of training protocol selected, which can range from performing a very low number of repetitions (3–5) with a very heavy weight and long rest periods to performing a very high number of repetitions (15–25) with a light weight and short rest periods, affects the magnitude of the hormonal responses (41,42) and thus probably affects the fuel used by the body for the activity. To date, acute changes in blood glucose during resistance exercise have only been measured for resistance exercise protocols involving three sets of eight repetitions at the maximum weight that can be lifted eight times safely with good form (43)—referred to as “8 RM,” or 8 repetition maximum (22,23,44). This regimen is at an intensity appropriate for “general conditioning” (43) and corresponds to a moderately high intensity.

Resistance exercise at 8 RM on its own results in modest declines in blood glucose in comparison to a similar duration of aerobic exercise in individuals with type 1 diabetes (22). Performing resistance exercise before aerobic exercise has been shown to have a protective effect against hypoglycemia (23,44). In one of our studies (44), when 12 participants with well-controlled type 1 diabetes who were regular exercisers performed 45 minutes of resistance exercise before 45 minutes of aerobic exercise, declines in blood glucose during the aerobic portion of the exercise session were attenuated (from 166 ± 61 to 105 ± 36 mg/dL, P = 0.001 without prior resistance exercise and from 166 ± 72 to 124 ± 56 mg/dL, P = 0.04 with prior resistance exercise). Performing resistance exercise after aerobic exercise can also prevent further declines in blood glucose levels (23).

Although the reasons for the protective effect of resistance exercise have not been fully elucidated, it is likely that hormonal responses play an important role. Performing resistance exercise before aerobic exercise has been associated with higher growth hormone levels than when aerobic exercise is performed first (40), which may promote a greater reliance on lipolysis during the aerobic portion of the exercise session, thereby sparing blood glucose. It is also possible that resistance exercise stimulates the release of epinephrine (45), thereby allowing greater access to both muscular and hepatic glycogen as a fuel source during exercise. Finally, the lactate produced during the performance of resistance exercise may assist in stabilizing subsequent blood glucose levels via production of new glucose through the Cori cycle in the liver.

Overall, the few studies of resistance exercise performed thus far seem to indicate that this form of physical activity may produce less of a disturbance in blood glucose levels during exercise and may attenuate declines in blood glucose when aerobic exercise is performed immediately subsequent to it. Further studies of different intensities and durations of resistance exercise, as well as different combinations of aerobic and resistance exercise (e.g., circuit training) will help shed more light on the utility of resistance exercise in the prevention of exercise-related hypoglycemia in individuals with type 1 diabetes.

It is important to note that the majority of the studies described in this review have performed their testing on mostly young (<30 years of age), physically active individuals. In addition, most of the participants in these studies were male. Both of these points are important: the factors (mostly hormones) that have been suggested as providing sprints, high-intensity intervals, and resistance exercise with their protective effect against declines in blood glucose during exercise vary with the age (46) and sex (46,47) of participants. It is also possible that responses vary depending on the fitness level (26) of the people performing the exercise. Finally, most of the studies involved cycling exercise, which has different energy requirements and potentially different hormonal responses than other forms of activity such as running or swimming. Further studies varying the types of exercise performed are required, as are more studies involving females, older individuals, and participants of different fitness levels to determine how these variations affect the magnitude of the protective effects of anaerobic exercise against declines in blood glucose in individuals with type 1 diabetes. Until then, frequent blood glucose testing or continuous glucose monitoring (CGM) should be recommended until individuals have ascertained their own responses to different exercise modalities.

Individuals who exercise regularly will eventually recognize that many factors influence their performance levels and that responses to exercise can be extremely variable. Some of these factors include sleep, hydration, immune system responses, and prior exercise. Where blood glucose responses to high-intensity exercise are concerned, prior exercise is likely to have a substantial impact. Performing a prior bout of either very intense or very protracted exercise can deplete glycogen stores in the body. As mentioned above, when access to glucose stores is limited, it is likely that the protective effects of high-intensity exercise will be attenuated because these rely to a great extent on glucose stored in the liver. We are not aware of any published studies examining the effect of glycogen depletion on blood glucose responses to sprints, resistance exercise, or high-intensity, intermittent exercise in individuals with type 1 diabetes.

The risk of post-exercise hypoglycemia in type 1 diabetes, particularly at night, has varied among studies. Prolonged exercise can result in depletion of glycogen stores, and repletion of these stores requires increasing glucose uptake from the blood. If insufficient carbohydrate is consumed after exercise, the risk of late-onset, post-exercise hypoglycemia may be increased.

Four studies to date have used CGM to examine the overnight period after aerobic exercise with a 10-second sprint or after intermittent, high-intensity exercise and have yielded mixed results (17,28–30). One study following participants for 8 hours after exercise did not find any effect of adding a 10-second sprint to the end of a moderate (40% VO₂peak) aerobic exercise session on the glucose requirements to prevent late-onset hypoglycemia (compared to aerobic exercise on its own) (17), indicating that the protective effects of sprinting may not extend far beyond the 2-hour window that has been examined to date. In another study, Maran et al. (28) found that blood glucose levels were significantly lower between midnight and 6:00 a.m. the night after an intermittent, high-intensity exercise session compared to a moderate exercise session. On the other hand, Iscoe and Riddell (29) found that interstitial glucose levels measured by CGM overnight were higher after an intermittent, high-intensity exercise session compared to moderate exercise alone. Campbell et al. (30), meanwhile, found no difference in nocturnal interstitial glucose levels between a day when intermittent, high-intensity activity was performed and a day when moderate aerobic exercise was performed.

Several factors could be responsible for the conflicting results of the latter three studies. The studies by Maran et al. (28) and Iscoe and Riddell (29) both involved cycling exercise, whereas that of Campbell et al. (30) involved running. Participants in the Maran et al. study (28) performed shorter (5 seconds compared to 15 seconds) and less intense (85% VO₂peak compared to >90% VO₂peak) intervals than those in the study by Iscoe and Riddell (29), which could have diminished the counterregulatory responses associated with high-intensity exercise. In addition, participants in the Iscoe and Riddell study (29) had higher aerobic fitness (42.4 ± 1.6 mL/kg/min) than the participants in the study by Maran et al. (28) (33.6 ± 6.1 mL/kg/min), allowing them to exercise at a higher absolute intensity.

Current studies examining the late post-exercise/nocturnal effects of intermittent, high-intensity exercise are not conclusive in determining whether the risk of hypoglycemia is increased or decreased by this type of activity because of a great deal of variability in both the protocols and the measured outcomes. Once again, this underscores the need for more research that can tease out the various factors at play, the magnitude of their influence, and the degree to which different physiological factors (e.g., age, sex, and fitness level) affect them.

Research to date gives us reason to be optimistic that anaerobic activities such as short sprints, high-intensity intervals, and resistance exercise can all be used successfully to help prevent hypoglycemia during exercise and thus decrease the reliance on excessive carbohydrate consumption during exercise to maintain blood glucose levels. It is, nonetheless, important for individuals to be able to understand and predict their own responses to these types of activities based on past experiences of trial and error. Because many physical, emotional, and physiological factors influence blood glucose responses to exercise on any given day, it still remains important to monitor blood glucose levels before, during, and after exercise to ensure an adequate level of safety.

While writing this manuscript, J.E.Y. was supported by a postdoctoral fellowship from the Canadian Institutes of Health Research. R.J.S. was supported by a Health Senior Scholar award from Alberta Innovates-Health Solutions.

No potential conflicts of interest relevant to this article were reported.