Abstract
Aims:
To evaluate factors affecting within-participant reproducibility in glycemic response to different forms of exercise.
Methods:
Structured exercise sessions ~30 minutes in length from the Type 1 Diabetes Exercise Initiative (T1DEXI) study were used to assess within-participant glycemic variability during and after exercise. The effect of several pre-exercise factors on the within-participant glycemic variability was evaluated.
Results:
Data from 476 adults with type 1 diabetes were analyzed. A participant’s change in glucose during exercise was reproducible within 15 mg/dL of the participant’s other exercise sessions only 32% of the time. Participants who exercised with lower and more consistent glucose level, insulin on board (IOB), and carbohydrate intake at exercise start had less variability in glycemic change during exercise. Participants with lower mean glucose (P < .001), lower glucose coefficient of variation (CV) (P < .001), and lower % time <70 mg/dL (P = .005) on sedentary days had less variable 24-hour post-exercise mean glucose.
Conclusions:
Reproducibility of change in glucose during exercise was low in this cohort of adults with T1D, but more consistency in pre-exercise glucose levels, IOB, and carbohydrates may increase this reproducibility. Mean glucose variability in the 24 hours after exercise is influenced more by the participant’s overall glycemic control than other modifiable factors.
Keywords: type 1 diabetes, exercise, reproducibility, glucose monitoring, remote food photography
Introduction
There are barriers to physical activity for people living with type 1 diabetes, sometimes due to unpredictable glycemic changes during exercise and fear of exercise-induced hypoglycemia.1,2 For instance, different forms of exercise (aerobic, resistance, mixed) result in different patterns of glycemic responses in type 1 diabetes, both during and after the activity.3,4 In general, prolonged light to moderate-intensity aerobic exercise activities, such as running or cycling, result in a variable drop in glycemia in persons with type 1 diabetes, 5 whereas brief bursts of aerobic/anerobic work6,7 and resistance exercise8,9 can result in a rise in glucose level, particularly if done in a fasted state. 10 Intermittent high-intensity exercise sessions intermixed with periods of light to moderate work can also promote a rise in glucose when fasted11,12 or have a moderating effect on glycemia if done at other times of the day.4,12-15 In general, prolonged bouts of physical activities tend to increase overnight hypoglycemia risk, particularly if the activity is later in the day.16,17 Exercise increases insulin sensitivity post-exercise and tends to block the counter-regulatory response to hypoglycemia, particularly if hypoglycemia occurs during the activity.18,19
Within each exercise type, considerable interindividual variation in the glycemic responses to exercise exists; 20 however, within individuals, some reproducibility does appear to exist, particularly if the exercise is performed “in laboratory” and in a fasted state.12,21-23 Even under conditions of highly repeatable laboratory conditions where identical meals and exercises are done at across days at identical times, glucose response can vary substantially during exercise within individuals. 24 Within-individual variation in the glucose response to exercise appears to increase if the conditions are less standardized, such as pre-meal vs post-meal or if the exercise is done at different times of the day.25,26 The purpose of this exploratory analysis was to characterize reproducibility of change in glucose during repeated bouts of three different forms of structured exercise (aerobic, resistance, interval) and examine how much of the within-individual variation in this glycemic response can be explained by various factors in the Type 1 Diabetes and Exercise Initiative (T1DEXI) study cohort, a large, real-world study of at home exercise of adults living with type 1 diabetes.
Methods
Study Design and Population
The T1DEXI study was designed to collect a variety of data surrounding exercise for individuals with type 1 diabetes. The study has been described elsewhere 20 and summarized herein. The protocol was approved by an Institutional Review Board, and each participant provided informed consent prior to starting the study. Adults (≥18 years of age), with a minimum two-year duration of type 1 diabetes, who were using either a commercially approved hybrid closed-loop (HCL) system, a standard insulin pump, or multiple daily injections (MDIs) to administer insulin were randomly assigned to complete one of three types of exercise videos designed for the study: aerobic, interval, or resistance. Each study-designed exercise video was ~30 minutes in duration, and participants were instructed to complete at least six sessions at home over a four-week period. Aerobic and interval exercise targeted 70% to 80% and 80% to 90% of the age-predicted maximal heart rate, respectively, whereas resistance exercise targeted major muscle group fatigue after three sets of eight resistance-band repetitions. Participants also continued their typical forms of physical activity, with a goal of completing 150 minutes per week of exercise which included study exercise videos and typical activity, and used a study-developed, cloud-connected smartphone application to enter information about exercise. Food intake, including pre-meal and post-meal food photos, was also self-reported through the T1DEXI smartphone app. 27 The Remote Food Photography Method (RFPM) was used to measure energy and nutrient intake from the food photos, as previously described. 27 Diabetes history, HbA1c, insulin delivery modality, insulin type, and other patient demographics were self-reported and collected via an online portal. Participants used their personal Dexcom continuous glucose monitor (CGM), or a blinded Dexcom G6 CGM San Diego, California if they did not use a personal Dexcom CGM, and a Verily Study Watch South San Francisco, California to collect continuous glucose and heart rate data throughout the four-week study period. Insulin delivery data were extracted using a diabetes data digital platform (Tidepool, Palo Alto, California) for those on continuous subcutaneous insulin infusion (CSII), whereas the T1DEXI smartphone app and/or a Mallya device (Biocorp, Issoire, France) was used to collect insulin data for those on multiple daily insulin injections.
Statistical Methods
This analysis included the structured exercise sessions involving the T1DEXI study-designed exercise videos, which participants were asked to repeat at least six times during the study. Participants with at least three exercise sessions with CGM data were included in the analyses. Bolus insulin on board (IOB), an estimate of the amount of active insulin in the body after a bolus dose for a meal, snack, or correction for hyperglycemia, 28 was estimated using a 4-hour linear decay model (ie, an administered bolus is reduced linearly to zero units after 4 hours). Participant-level reproducibility of change in glucose during exercise was assessed by computing the within-participant standard deviation (SD) of change in glucose during exercise, with lower values indicating more consistent (reproducible) during-exercise glucose response. The percentage of exercise sessions where glycemic change during exercise is within 5, 10, and 15 mg/dL of the glycemic change in the other exercise sessions for each participant was calculated to further describe reproducibility of glycemic response to exercise. The association of participant-level factors with within-participant SD of change in glucose during exercise was assessed through a linear regression model adjusting for baseline HbA1c, age, average glucose level at the start of exercise, insulin delivery modality, and sex.
Within-participant SD of mean glucose in the 24 hours after exercise was used to measure consistency of post-exercise mean glucose control. Post-exercise periods were truncated if another study video or personal exercise session was reported within the 24-hour period. Mean glucose was calculated for post-exercise periods with ≥20 hours of CGM data in each period and the within-participant SD of mean glucose was analyzed for participants with ≥3 post-exercise periods that met minimum data requirements. The association of participant-level factors with within-participant SD of post-exercise mean glucose was assessed through a linear regression model adjusting for baseline HbA1c, age, mean glucose on sedentary days, insulin modality, and sex. These covariates were selected due to their potential association with the participants’ glycemic control.
Continuous factors were tested as a continuous predictor in each model, but for table display purposes, continuous factors were categorized into groups with similar counts to demonstrate the relationship between each continuous factor level vs the within-participant SD of change in glucose during exercise or within-participant SD of mean glucose 24 hours after exercise. All statistical tests were two-sided. Multiple comparisons were corrected using the Benjamini-Hochberg adaptive false discovery rate correction procedure. 29 Analyses were performed with SAS software, version 9.4 (SAS Institute, Cary, North Carolina).
Results
Participant Characteristics
The study included 476 adults with type 1 diabetes; mean (range) age of 34 (18-70) years and HbA1c of 6.6% ± 0.7% (49 ± 7.7 mmol/mol) (mean ± SD). Ninety-five percent of participants were using CGM at baseline, and 45% of participants were using an HCL system. Additional baseline characteristics are displayed in Table 1.
Table 1.
Participant Characteristics at Baseline.
| Overall (N = 476) | |
|---|---|
| Age (yrs) median (quartiles) | 34 (26, 46) |
| mean (SD) | 37 (14) |
| 18 to 24 n (%) | 98 (21%) |
| 25 to 44 n (%) | 246 (52%) |
| 45 to 59 n (%) | 83 (17%) |
| ≥60 n (%) | 49 (10%) |
| range (min to max) | 18 to 70 |
| Gender—female (%) | 347 (73%) |
| HbA1c—% [mmol/mol] mean (SD) | 6.6 (0.7) [49±7.7] |
| <6.0 [<42 mmol/mol] n (%) | 85 (18%) |
| 6.0 to 6.4 [42 to 46 mmol/mol] n (%) | 122 (26%) |
| 6.5 to 6.9 [48 to 52 mmol/mol] n (%) | 123 (26%) |
| 7.0 to 7.4 [53 to 57 mmol/mol] n (%) | 83 (17%) |
| 7.5 to 7.9 [58 to 63 mmol/mol] n (%) | 39 (8%) |
| ≥8.0 [≥64 mmol/mol] n (%) | 24 (5%) |
| Range (min to max) | 4.8 to 10.0 [29 to 86] |
| Type 1 Diabetes Duration (yrs) | |
| Median (IQR) | 15 (9, 25) |
| Range | 2 to 62 |
| Body-Mass Index (kg/m2) mean (SD) | 25.3 (4.0) |
| Underweight (<18.5) n (%) | 1 (<1%) |
| Normal (18.5 to <25.0) n (%) | 263 (55%) |
| Overweight (25.0 to <30.0) n (%) | 151 (32%) |
| Obese (≥30.0) n (%) | 61 (13%) |
| Range (min to max) | 18.2 to 42.0 |
| Race/ethnicity n (%) | |
| White Non-Hispanic | 425 (89%) |
| Black/African American | 8 (2%) |
| Hispanic or Latino | 13 (3%) |
| Asian | 9 (2%) |
| American Indian/Alaskan Native | 2 (<1%) |
| More than one race | 8 (2%) |
| Unknown/not reported | 11 (2%) |
| Highest Education Level n (%) | |
| <Bachelor’s | 86 (18%) |
| Bachelor’s | 222 (47%) |
| >Bachelor’s | 166 (35%) |
| Unknown/not reported | 2 (<1%) |
| Annual income n (%) | |
| <$50 000 | 69 (14%) |
| $50 000 to <$100 000 | 142 (30%) |
| ≥$100 000 | 196 (41%) |
| Unknown/not reported | 69 (14%) |
| Insurance n (%) | |
| Private | 408 (86%) |
| Other | 45 (9%) |
| None | 2 (<1%) |
| Unknown/not reported | 21 (4%) |
| Insulin modality at enrollment n (%) | |
| Closed loop | 213 (45%) |
| Pump | 83 (17%) |
| Injections | 180 (38%) |
| Current CGM User n (%) | 453 (95%) |
| CGM Device Type (for participants using CGM) n (%) | N = 453 |
| Abbott | 11 (2%) |
| Dexcom | 401 (89%) |
| Medtronic | 41 (9%) |
| Most recent severe hypoglycemic event n (%)a | |
| Never | 256 (54%) |
| <1 year ago | 22 (5%) |
| ≥1 year ago | 198 (42%) |
| Most recent DKA event n (%)b | |
| Never | 296 (62%) |
| <1 year ago | 34 (7%) |
| ≥1 year ago | 146 (31%) |
| International Physical Activity Questionnaire (IPAQ) Category n (%) | N = 448 |
| Inactivec | 35 (8%) |
| Minimally actived | 224 (50%) |
| HEPA activee | 189 (42%) |
| Clarke Hypoglycemia Awareness Survey Category n (%) | |
| Aware | 316 (67%) |
| Uncertain | 67 (14%) |
| Reduced Awareness | 91 (19%) |
| Pittsburg Sleep Quality Index (PSQI) Scoref median (IQR) | 5 (3, 7) |
Abbreviations: SD, standard deviation; IQR, interquartile range; CGM, continuous glucose monitor; DKA, diabetic ketoacidosis; IPAQ, International Physical Activity Questionnaire; HEPA, health-enhancing physical activity; PSQI, Pittsburg Sleep Quality Index.
During-Exercise Glycemic Change
The average age-adjusted mean heart rate reserve during aerobic, interval, and resistance exercise was 39% ± 14%, 29% ± 14%, and 27% ± 11%, respectively. Participants reported consuming a snack during 6% of the exercise sessions. Additional information on temporary basal rates and pump suspensions prior to and during exercise has been previously published elsewhere. 20 Average glucose at the start of exercise was 149 ± 31 mg/dL and average change in glucose during exercise was −13 ± 19 mg/dL. The average within-participant SD of change in glucose during exercise (ie, the measure of reproducibility of change in glucose during exercise) was 30 ± 15 mg/dL. Twelve percent of exercises had a change in glucose within 5 mg/dL of the change in glucose during the participant’s other exercise sessions, 22% of exercises within 10 mg/dL, and 32% of exercises within 15 mg/dL. For participants with a lower within-participant change in glucose during exercise SD of <20 mg/dL, the percentage of exercise sessions with a change in glucose within 15 mg/dL of their other exercise sessions was 53%; this decreased to 31% for participants with a within-participant SD of change in glucose of 20 to <30 mg/dL, 23% for an SD of 30 to <40 mg/dL, and 15% for an SD of ≥40 mg/dL (Table 2).
Table 2.
Reproducibility of Change in Glucose During Exercise Within Each Participant.
| Number of participants | Number of exercise sessions per participant, Mean (SD) | Percent of exercise sessions where change in glucose during exercise was within | |||
|---|---|---|---|---|---|
| ±5 mg/dL of each other, % | ±10 mg/dL of each other, % | ±15 mg/dL of each other, % | |||
| Overall | 476 | 6 (1) | 12 | 22 | 32 |
| Within-participant SD of change in glucose during exercise | |||||
| <20 mg/dL | 131 | 6 (1) | 20 | 37 | 53 |
| 20 to <30 mg/dL | 137 | 6 (1) | 11 | 21 | 31 |
| 30 to <40 mg/dL | 99 | 6 (1) | 9 | 17 | 23 |
| ≥40 mg/dL | 109 | 6 (1) | 5 | 10 | 15 |
Abbreviation: SD, standard deviation.
Lower average glucose at start of exercise, lower average IOB at the start of exercise, and lower average carbohydrate consumption in the hour prior to structured exercise were associated with a more consistent change in glucose during structured exercise. Specifically, the within-participant SD of change in glucose during exercise was lower (ie, change in glucose during exercise was more consistent) for participants with a lower average glucose at start of exercise (within-participant change in glucose SD of 22 ± 12 vs 35 ± 17 mg/dL for those with an average starting glucose of <130 vs ≥160 mg/dL, P < .001); for those with a lower average IOB at the start of exercise (SD of 25 ± 14 vs 35 ± 16 mg/dL for those with an average IOB at the start of exercise of <1 vs ≥2 units, P = .003); and for those with a lower average carbohydrate consumption in the hour prior to exercise (SD of 30 ± 15 vs 35 ± 17 mg/dL for those that consumed an average of >0 to <5 vs ≥15 g carbohydrates in the hour prior to exercise, P < .001; Table 3).
Table 3.
Association of Participant-Level Factors vs. Within-Participant SD of Change in Glucose During Exercise.
| Characteristic | Number of participants | Number of exercise sessions | Average glucose at start of exercise (mg/dL), Mean (SD) | Average change in glucose during exercise (mg/dL), Mean (SD) | Within-participant SD of change in glucose during exercise (mg/dL), Mean (SD) | P-value a |
|---|---|---|---|---|---|---|
| Overall | 476 | 2719 | 149 (31) | -13 (19) | 30 (15) | |
| Baseline age (yrs) | .01 | |||||
| 18 to 25 | 98 | 533 | 159 (37) | -15 (20) | 34 (16) | |
| 26 to 44 | 246 | 1404 | 148 (30) | -13 (19) | 30 (16) | |
| ≥45 | 132 | 782 | 145 (27) | -13 (17) | 27 (13) | |
| T1D duration (yrs) | .31 | |||||
| <5 | 56 | 311 | 137 (29) | -6 (20) | 26 (13) | |
| 5 to <10 | 78 | 441 | 152 (37) | -13 (18) | 30 (16) | |
| ≥10 | 342 | 1967 | 151 (30) | -14 (18) | 31 (15) | |
| Baseline HbA1c (%) | .79 | |||||
| <6.0 [<42 mmol/mol] | 85 | 486 | 133 (24) | -11 (18) | 27 (15) | |
| 6.0 to 6.4 [42 to 46 mmol/mol] | 122 | 694 | 140 (26) | -13 (21) | 28 (14) | |
| 6.5 to 6.9 [48 to 52 mmol/mol] | 123 | 709 | 148 (26) | -12 (17) | 31 (15) | |
| 7.0 to 7.4 [53 to 57 mmol/mol] | 83 | 481 | 161 (28) | -16 (17) | 33 (15) | |
| 7.5 to 7.9 [58 to 63 mmol/mol] | 39 | 214 | 174 (41) | -16 (20) | 32 (20) | |
| ≥8.0 [≥64 mmol/mol] | 24 | 135 | 179 (38) | -17 (14) | 35 (16) | |
| Baseline BMI (kg/m2) | .05 | |||||
| <25.0 | 264 | 1502 | 148 (31) | -11 (19) | 31 (16) | |
| 25.0 to <30.0 | 151 | 866 | 153 (30) | -18 (17) | 31 (14) | |
| ≥30.0 | 61 | 351 | 147 (35) | -13 (17) | 26 (13) | |
| Within-participant SD of glucose at start of exercise (mg/dL) | <.001 | |||||
| <25 | 103 | 576 | 124 (21) | -6 (17) | 21 (11) | |
| 25 to <50 | 235 | 1365 | 148 (25) | -14 (18) | 29 (13) | |
| ≥50 | 138 | 778 | 171 (32) | -17 (20) | 39 (17) | |
| Average glucose at start of exercise (mg/dL) | <.001 | |||||
| <130 | 129 | 725 | 115 (11) | -3 (15) | 22 (12) | |
| 130 to <160 | 197 | 1145 | 145 (9) | -14 (18) | 32 (14) | |
| ≥160 | 150 | 849 | 185 (24) | -22 (18) | 35 (17) | |
| Within-participant SD of exercise time of day (hr) | .38 | |||||
| <1 | 94 | 525 | 147 (29) | -15 (20) | 28 (16) | |
| 1 to <3 | 200 | 1143 | 152 (33) | -13 (18) | 32 (16) | |
| ≥3 | 182 | 1051 | 148 (30) | -13 (18) | 30 (15) | |
| Average insulin on board at the start of exercise (units) | .003 | |||||
| <1 | 141 | 816 | 139 (31) | -5 (15) | 25 (14) | |
| 1 to <2 | 145 | 832 | 150 (30) | -15 (18) | 30 (15) | |
| ≥2 | 178 | 1009 | 158 (30) | -19 (20) | 35 (16) | |
| Within-participant SD of insulin on board at the start of exercise (units) | .02 | |||||
| <1 | 194 | 1113 | 141 (28) | -8 (17) | 27 (14) | |
| 1 to <2 | 157 | 916 | 150 (28) | -16 (19) | 32 (15) | |
| ≥2 | 113 | 628 | 165 (34) | -19 (18) | 35 (17) | |
| Resting heart rate after wake (beats/min) | .77 | |||||
| <70 | 175 | 1012 | 142 (28) | -12 (18) | 29 (15) | |
| 70 to <80 | 169 | 970 | 153 (32) | -14 (18) | 30 (15) | |
| ≥80 | 123 | 692 | 153 (32) | -13 (20) | 33 (17) | |
| Average carbs consumed One hour before exercise (g) | <.001 | |||||
| 0 | 129 | 735 | 149 (31) | -14 (20) | 27 (14) | |
| >0 to <5 | 57 | 338 | 144 (26) | -13 (17) | 30 (15) | |
| 5 to <15 | 87 | 508 | 153 (25) | -14 (18) | 36 (14) | |
| ≥15 | 57 | 328 | 151 (35) | -14 (22) | 35 (17) | |
| Within-participant SD of carbs consumed One hour before exercise (g) | <.001 | |||||
| <5 | 153 | 877 | 148 (31) | -14 (19) | 27 (13) | |
| 5 to <15 | 81 | 471 | 148 (28) | -12 (18) | 34 (16) | |
| ≥15 | 96 | 561 | 153 (30) | -15 (20) | 36 (15) | |
| Average mean daily carb intake (g) during past Two weeks | .42 | |||||
| <100 | 17 | 105 | 130 (23) | -9 (14) | 21 (11) | |
| 100 to <150 | 77 | 452 | 151 (28) | -14 (18) | 32 (15) | |
| 150 to <200 | 80 | 488 | 147 (24) | -16 (20) | 31 (14) | |
| ≥200 | 74 | 434 | 151 (31) | -13 (21) | 31 (15) | |
| Within-participant SD of mean daily carb intake (g) during past Two weeks | .29 | |||||
| <10 | 97 | 579 | 146 (27) | -13 (21) | 29 (15) | |
| 10 to <20 | 79 | 477 | 147 (26) | -16 (20) | 31 (13) | |
| ≥20 | 72 | 423 | 153 (30) | -14 (17) | 33 (15) | |
| Sex | .83 | |||||
| Female | 346 | 1992 | 152 (31) | -13 (18) | 31 (15) | |
| Male | 130 | 727 | 142 (31) | -15 (19) | 29 (15) | |
| Insulin modality | .09 | |||||
| Closed loop | 213 | 1225 | 149 (28) | -13 (20) | 31 (15) | |
| MDI | 83 | 475 | 153 (39) | -13 (16) | 28 (19) | |
| Pump | 180 | 1019 | 148 (31) | -13 (18) | 30 (14) | |
| Exercise type | .03 | |||||
| Aerobic | 157 | 897 | 150 (30) | -18 (21) | 32 (16) | |
| Interval | 156 | 887 | 149 (31) | -14 (16) | 28 (14) | |
| Resistance | 163 | 935 | 149 (32) | -8 (17) | 31 (16) |
Abbreviations: SD, standard deviation; BMI, body mass index; MDI, multiple daily injection.
From a linear regression model adjusting for baseline HbA1c, baseline age, average glucose at the start of exercise, insulin modality, and sex. Multiple comparisons were adjusted for using the adaptive Benjamini-Hochberg false discovery rate correction procedure.
Not only were lower average starting glucose, IOB, and carbohydrate consumption associated with a more reproducible change in glucose during exercise but maintaining similar glycemic levels (P < .001), IOB (P = .02), and carbohydrate consumption (P < .001) at the start of each exercise was also associated with a more reproducible glycemic effect during structured exercise (Table 3). The average pre-exercise levels of glucose, insulin, and carbohydrate consumption that characterize different levels of reproducibility of change in glucose during exercise are summarized in Supplemental Table 1. However, average daily carbohydrate consumption in the two weeks prior to exercise was not associated with a more reproducible glycemic response during exercise (P = .42, Table 3). Scatterplots and boxplots of the association between these three factors and within-participant SD of change in glucose during exercise are displayed in Figure 1.
Figure 1.
Relationship between within-participant SD of change in glucose during exercise and exercise factors. Scatterplots of within-participant SD of change in glucose during exercise vs (a) mean glucose at start of exercise (n = 476), (b) within-participant SD of glucose at start of exercise (N = 476), (c) mean insulin on board at the start of exercise (N = 464), and (d) within-participant SD of insulin on board prior to exercise (N = 464), one point per participant. Boxplots of within-participant SD of change in glucose during exercise vs (e) mean carbohydrates consumed within the hour prior to exercise (N = 330), and (f) within-participant SD of carbohydrates consumed within the hour prior to exercise (N = 330).
Numbers on top of boxes represent number of participants, dots in the middle of boxes indicate means, lines in the middle of boxes indicate medians, and the bottom and top of boxes represent the 25th and 75th percentiles, respectively.
Change in glucose during exercise was also more consistent for participants assigned to interval exercise when compared with resistance and aerobic exercise (within-participant change in glucose SD of 28 ± 14 vs 32 ± 16 and 31 ± 16 mg/dL for aerobic and resistance exercise, respectively, P = .03, Table 3).
Post-exercise Glycemic Control
Average mean glucose in the 24 hours post-exercise was 144 ± 24 mg/dL. Average within-participant SD of post-exercise mean glucose was 18 ± 13 mg/dL. Post-exercise mean glucose was more consistent for participants with lower sedentary day mean glucose (13 ± 7 vs 26 ± 19 mg/dL for participants with sedentary day mean glucose of <140 vs ≥160 mg/dL, P < .001); for those with lower sedentary day glucose CV (14 ± 8 vs 26 ± 18 mg/dL for participants with sedentary day glucose CV of <30% vs ≥35%, P < .001); for those with a lower sedentary day time <70 mg/dL (17 ± 9 vs 21 ± 16 mg/dL for participants with sedentary day time <70 mg/dL of <1% vs ≥2%, P = .005); for those with lower within-participant SD of time <70 mg/dL on sedentary days (17 ± 9 vs 23 ± 16 mg/dL for participants with sedentary day time <70 mg/dL within-participant SD of <1% vs ≥4%, P = .002); and for those using an HCL system (15 ± 9 mg/dL for HCL vs 19 ± 11 mg/dL for pump without automation vs 23 ± 16 mg/dL for MDI, P = .004; Table 4). Exercise type, within-participant SD of change in glucose, and within-participant SD of exercise time of day were not associated with a more consistent post-exercise mean glucose.
Table 4.
Association of Participant-Level Factors vs Within-Participant SD of Mean Glucose on Exercise Days.
| Characteristic | Number of participants | Mean glucose on exercise days (mg/dL), Mean (SD) | Within-participant SD on exercise days (mg/dL), Mean (SD) | P-value a |
|---|---|---|---|---|
| Overall | 232 | 144 (24) | 18 (13) | |
| Baseline age (yrs) | .87 | |||
| 18 to 25 | 56 | 153 (25) | 23 (16) | |
| 26 to 44 | 123 | 144 (26) | 17 (12) | |
| ≥45 | 53 | 136 (16) | 16 (9) | |
| T1D duration (yrs) | .62 | |||
| <5 | 22 | 131 (22) | 16 (10) | |
| 5 to <10 | 46 | 144 (22) | 21 (18) | |
| ≥10 | 164 | 146 (25) | 18 (12) | |
| Baseline HbA1c (%) | .23 | |||
| <6.0 [<42 mmol/mol] | 42 | 125 (13) | 12 (7) | |
| 6.0 to 6.4 [42 to 46 mmol/mol] | 58 | 139 (15) | 17 (10) | |
| 6.5 to 6.9 [48 to 52 mmol/mol] | 62 | 144 (19) | 18 (11) | |
| 7.0 to 7.4 [53 to 57 mmol/mol] | 40 | 153 (20) | 20 (10) | |
| 7.5 to 7.9 [58 to 63 mmol/mol] | 18 | 159 (25) | 24 (21) | |
| ≥8.0 [≥64 mmol/mol] | 12 | 192 (39) | 35 (22) | |
| Baseline BMI (kg/m2) | .30 | |||
| <25.0 | 118 | 146 (23) | 19 (14) | |
| 25.0 to <30.0 | 77 | 143 (24) | 17 (9) | |
| ≥30.0 | 37 | 143 (29) | 19 (16) | |
| Within-participant SD of exercise time of day (hr) | .50 | |||
| <1 | 54 | 145 (22) | 19 (10) | |
| 1 to <3 | 94 | 142 (22) | 18 (15) | |
| ≥3 | 84 | 147 (28) | 19 (12) | |
| Average insulin on board at the start of exercise (units) | .17 | |||
| <1 | 64 | 141 (25) | 16 (13) | |
| 1 to <2 | 67 | 146 (23) | 21 (14) | |
| ≥2 | 87 | 145 (25) | 17 (11) | |
| Within-participant SD of insulin on board at the start of exercise (units) | .77 | |||
| <1 | 86 | 140 (24) | 16 (12) | |
| 1 to <2 | 65 | 145 (23) | 18 (11) | |
| ≥2 | 67 | 149 (26) | 20 (14) | |
| Resting heart rate after wake (beats/min) | .86 | |||
| <70 | 62 | 138 (22) | 16 (10) | |
| 70 to <80 | 90 | 144 (21) | 18 (13) | |
| ≥80 | 78 | 150 (28) | 21 (15) | |
| Mean glucose (mg/dL) on sedentary days | <.001 | |||
| <140 | 88 | 126 (11) | 13 (7) | |
| 140 to <160 | 81 | 145 (12) | 18 (9) | |
| ≥160 | 63 | 170 (26) | 26 (19) | |
| Glucose CV (%) on sedentary days | <.001 | |||
| <30% | 94 | 136 (18) | 14 (8) | |
| 30 to <35% | 71 | 147 (23) | 17 (9) | |
| ≥35% | 67 | 153 (29) | 26 (18) | |
| % Time <70 mg/dL on sedentary days | .005 | |||
| <1% | 98 | 150 (25) | 17 (9) | |
| 1 to <2% | 49 | 144 (26) | 16 (13) | |
| ≥2% | 85 | 139 (22) | 21 (16) | |
| Within-participant SD of % time <70 mg/dL on sedentary days | .002 | |||
| <1% | 59 | 148 (24) | 17 (9) | |
| 1% to <2% | 66 | 148 (24) | 17 (10) | |
| 2% to <4% | 58 | 143 (26) | 18 (16) | |
| ≥4% | 49 | 137 (21) | 23 (16) | |
| Within-participant SD of change in glucose during exercise (mg/dL) | .67 | |||
| <25 | 110 | 141 (20) | 17 (10) | |
| 25 to <50 | 91 | 143 (23) | 19 (13) | |
| ≥50 | 31 | 160 (34) | 23 (19) | |
| Within-participant SD of glucose level at the end of exercise (mg/dL) | .36 | |||
| <25 | 72 | 135 (21) | 16 (13) | |
| 25 to <50 | 110 | 145 (22) | 18 (11) | |
| ≥50 | 50 | 157 (27) | 22 (16) | |
| Within-participant SD of mean age- and gender-adjusted heart rate reserve during exercise (%) | .23 | |||
| <5% | 65 | 147 (30) | 20 (17) | |
| 5 to <10% | 82 | 143 (22) | 17 (9) | |
| ≥10% | 52 | 145 (21) | 19 (11) | |
| Average mean daily carb intake (g) during past two weeks | .60 | |||
| <100 | 1 | 117 | 9 | |
| 100 to <150 | 21 | 140 (23) | 18 (10) | |
| 150 to <200 | 24 | 140 (21) | 15 (9) | |
| ≥200 | 21 | 141 (19) | 18 (12) | |
| Within-participant SD of mean daily carb intake (g) during past two weeks | .14 | |||
| <10 | 26 | 139 (24) | 17 (10) | |
| 10 to <20 | 16 | 147 (24) | 20 (15) | |
| ≥20 | 25 | 137 (14) | 13 (5) | |
| Sex | .60 | |||
| Female | 173 | 146 (25) | 19 (13) | |
| Male | 59 | 139 (21) | 16 (13) | |
| Insulin modality | .004 | |||
| Closed loop | 114 | 142 (20) | 15 (9) | |
| MDI | 34 | 142 (26) | 19 (11) | |
| Pump | 84 | 148 (28) | 23 (16) | |
| Exercise type | .50 | |||
| Aerobic | 71 | 143 (25) | 17 (10) | |
| Interval | 69 | 144 (22) | 19 (13) | |
| Resistance | 92 | 146 (25) | 19 (15) |
Abbreviations: SD, standard deviation; BMI, body mass index; MDI, multiple daily injection.
To be included in this table, at least three exercise days were required from each participant with at least 20 hours of CGM data on each exercise day.
From a linear regression model adjusting for baseline HbA1c, baseline age, mean glucose on sedentary days, insulin modality, and sex. Multiple comparisons were adjusted for using the adaptive Benjamini-Hochberg false discovery rate correction procedure.
Discussion
In a population of adults living with type 1 diabetes, reproducibility of change in glucose during exercise was moderately low, with only 12% of repeated exercises of the same structure type (ie, aerobic, resistance, interval) having a change in glucose during exercise within 5 mg/dL of each other, and only 32% having a change in glucose within 15 mg/dL of each other. This low-to-modest reproducibility in the glucose changes during exercise is consistent with the results from a smaller study by Notkin et al 26 where the change in glucose during three in-lab cycling sessions was variable in adults with type 1 diabetes, even after attempting to control for pre-exercise factors. However, in the T1DEXI cohort, we found considerable variation in reproducibility between participants, with some participants having higher reproducibility than others.
In a previous analysis for the T1DEXI study, 20 it was found that exercise-specific factors such as exercise type, glucose at the start of exercise, IOB at the start of exercise, and exercise time of day were significantly associated with the nominal change in glucose during an individual study video-led exercise session. This subsequent analysis suggests that control of some of those pre-exercise factors were also associated with a more reproducible change in glucose during exercise. In this free-living cohort of adults with type 1 diabetes, both lower starting glucose and similar starting glucose from exercise to exercise were associated with a more consistent change in glucose during exercise and a smaller nominal mean drop in glucose. The effect of lower starting glucose yielding consistently smaller drops in glucose during exercise compared with situations where baseline glucose was elevated confirms previous findings in laboratory settings 23 but could also potentially be explained by participants reacting to their glucose levels dropping toward what might have been perceived as being too low for their own preference from a hypoglycemia mitigation perspective. 30
Similar relationships with reproducibility of change during exercise were seen with IOB at the start of exercise. A lower average IOB level, as well as similar IOB at the start of each exercise session, was both individually associated with a more reproducible change in glucose during exercise. These findings are in line with studies suggesting that fasting exercise, where IOB is typically zero unless a correction dose is given because of morning hyperglycemia, results in greater reproducibility in the glycemic responses to exercise.12,21-23 A higher average IOB yielded a less consistent during-exercise glycemic effect. The reason why IOB level increases variability in the acute glucose response to exercise is unclear but may be related to how exercise might alter insulin kinetics, 31 or how exercise may unpredictably alter insulin sensitivity based on factors such as exercise type and exercise time of day.32,33 In our analyses, the use of an closed-loop system did not appear to significantly influence the reproducibility of the glucose response to exercise, but closed-loop users tended to have higher time in range on both exercise and sedentary days as compared with either MDI or standard pump therapy. 20
Multiple investigators have hypothesized that a low-carbohydrate diet <100 g per day would improve glycemic levels for people living with type 1 diabetes; however, it is unclear if this applies to individuals with type 1 diabetes who are routinely physically active and thus may require higher carbohydrate intake to preserve exercise performance and mitigate hypoglycemia risk. 34 In this analysis, there was insufficient evidence to suggest that average daily carbohydrate intake in the two weeks prior to exercise impacted change in glucose during exercise or mean glucose post-exercise. However, a lower carbohydrate intake within the hour prior to exercise, as well as a more consistent pattern of carbohydrate intake from exercise to exercise, was associated with a more reproducible change in glucose during exercise. It should also be noted, however, that carbohydrate loading just before or during exercise with high rates of carbohydrate intake during prolonged exercise is possible for individuals living with type 1 diabetes in endurance training and competition settings, as long as insulin delivery rates are sufficient to prevent hyperglycemia and may help facilitate performance. 35 In this cohort, participants with the most reproducible change in glucose during exercise (SD <20 mg/dL) had an average glucose level of 138 mg/dL and a median average IOB of 1.08 units at start of exercise with little or no carbohydrates reported recently. These real-world findings suggest that a given glycemic response to repeated exercise of the same structure is somewhat reproducible if certain event-level factors are controlled, such as bolus insulin levels, feeding patterns, and pre-exercise glucose concentrations.
With regard to exercise type, even though resistance exercise had the smallest mean drop in glucose during exercise, the change in glucose during interval exercise was significantly more consistent when compared with aerobic or resistance exercise. Thus, change in glucose during exercise may be more easily predicted for interval exercise, although the reproducibility was only marginally greater. These findings support a previous study suggesting that glucose control after high-intensity interval exercise shows some degree of reproducibility, especially if baseline glucose is consistent and IOB is low or zero. 12 If high-intensity exercise is done under less standardized conditions, this activity is considerably more inconsistent, both among and within individuals. 14
Beyond the acute glycemic effect during exercise, factors affecting variability in mean glucose in the 24 hours after exercise were also explored in this analysis. The factors that were the most significantly associated with a consistent mean glucose after exercise were also factors reflective of participants with better baseline glycemic levels (ie, those with lower sedentary day mean glucose levels, lower glucose CV, and less % time <70 mg/dL). With regard to the impact of insulin modality, participants on closed-loop insulin dosing devices also had the least variability in 24-hour post-exercise mean glucose, followed by MDI, and then non-closed loop insulin dosing devices, even after adjusting for mean glucose on sedentary days. These findings support the notion that closed-loop systems tend to improve glucose control post-exercise as compared to standard pump therapy or MDI. 5
Strengths of this analysis include free-living data reporting that may more appropriately reflect day-to-day participant activity, in contrast to a lab-based study where exercise conditions are controlled. However, there are also notable limitations of this analysis. The study cohort is likely not representative of the general population living with type 1 diabetes, primarily because despite requiring that participants only be willing to engage in physical activity with a goal of at least 150 minutes of physical activity per week, the baseline physical activity patterns of the enrolled cohort were high and they had better glycemic control than the typical type 1 diabetes population. In addition, to reduce variability in change in glucose during exercise due to differences in exercise type, the analysis was limited to only the structured study-assigned video-based exercise sessions and did not explore the participants’ other elective exercise types, the latter of which may be of a longer duration and/or of a higher intensity. Finally, it is possible that other factors not explored in this article, such as reduced insulin or temporary basal rates, could have an effect on the reproducibility of change in glucose.
In summary, in free-living adults with type 1 diabetes, reproducibility of change in glucose during exercise was generally low, but this analysis also identified a limited number of factors that appear to be associated with a more reproducible glycemic response during and after exercise events. Specifically, a more consistent and lower starting glucose, a lower IOB, and a modest and reproducible carbohydrate intake level in the hour prior to exercise are all associated with a more reproducible in-exercise glycemic effect. In addition, it appears that having better glycemic control on sedentary days and the usage of a closed-loop insulin pump are also associated with a more stable post-exercise glycemia. These results may help inform new decision-support tools, including better guidance on exercise preparedness, for people living with type 1 diabetes who have difficulty managing glycemia around exercise.
Supplemental Material
Supplemental material, sj-docx-1-dst-10.1177_19322968241234687 for Factors Affecting Reproducibility of Change in Glucose During Exercise: Results From the Type 1 Diabetes and EXercise Initiative by Zoey Li, Peter Calhoun, Michael R. Rickels, Robin L. Gal, Roy W. Beck, Peter G. Jacobs, Mark A. Clements, Susana R. Patton, Jessica R. Castle, Corby K. Martin, Melanie B. Gillingham, Francis J. Doyle and Michael C. Riddell in Journal of Diabetes Science and Technology
Footnotes
Abbreviations: CGM, continuous glucose monitor; HCL, hybrid closed loop; IOB, insulin on board; MDI, multiple daily injection; RFPM, Remote Food Photography Method; SD, standard deviation; T1DEXI, Type 1 Diabetes Exercise Initiative Study.
The author(s) declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article: Z.L. reports no conflict of interests. M.R.R. reports consultancy fees from Zealand Pharma. R.L.G. reports no conflict of interests. P.C. reports no conflict of interests. P.G.J. reports receiving grants from the National Institutes of Health, The Leona M. and Harry B. Charitable Trust, the Juvenile Diabetes Research Foundation, Dexcom, and the Oregon Health & Science University Foundation; consultancy fees from CDISC; US patents 62/352,939, 63/269,094, 62/944,287, 8810388, 9,480,418, 8,317,700, 61/570382, 8,810,388, 7,976,466, and 6,558,321; and reports stock options from Pacific Diabetes Technologies, outside submitted work. M.A.C. is Chief Medical Officer of Glooko, Inc and has received grants or contracts from Dexcom, Abbott Diabetes Care, National Institutes of Health, the Juvenile Diabetes Research Foundation, the Emily Rosebud Foundation, Eli Lilly, Tolerion, and Garmin. F.J.D. reports no conflict of interests. S.R.P. reports receiving grants from The Leona M. and Harry B. Helmsley Charitable Trust, the National Institutes of Health, and the Jaeb Center for Health Research and honorarium from the American Diabetes Association, outside the submitted work. J.R.C. reports receiving grants from the Juvenile Diabetes Research Foundation, the National Institutes of Health, Dexcom, and Medtronic and consultancy fees from Novo Nordisk, Insulet, and Zealand, outside the submitted work. M.B.G. reports no conflict of interest. R.W.B. reports receiving consulting fees, paid to his institution, from Insulet, Bigfoot Biomedical, vTv Therapeutics, and Eli Lilly, grant support and supplies, provided to his institution, from Tandem and Dexcom, and supplies from Ascenia and Roche. C.K.M. reports no conflict of interests. M.C.R. reports receiving consulting fees from the Jaeb Center for Health Research, Eli Lilly, Zealand Pharma, and Zucara Therapuetics; speaker fees from Sanofi Diabetes, Eli Lilly, Dexcom Canada, and Novo Nordisk; and stock options from Supersapiens and Zucara Therapeutics.
Funding: The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Research reported in this publication was supported by The Leona M. and Harry. B. Helmsley Charitable Trust. The content is solely the responsibility of the authors and does not necessarily represent the official views of The Leona M. and Harry B. Helmsley Charitable Trust. One of the author’s institutions (author C.K.M., Pennington Biomedical Research Center) is supported by NORC Center Grant P30 DK072476 entitled “Nutrition and Metabolic Health Through the Lifespan” sponsored by NIDDK and by grant U54 GM104940 from the National Institute of General Medical Sciences, which funds the Louisiana Clinical and Translational Science Center. Non-financial Support: Verily (South San Francisco, CA) provided the Study Watch at no cost. Dexcom provided continuous glucose monitors at a discounted rate.
ORCID iDs: Zoey Li 
https://orcid.org/0000-0002-5950-9317
Peter G. Jacobs
https://orcid.org/0000-0001-9897-4783
Mark A. Clements
https://orcid.org/0000-0002-2368-0331
Susana R. Patton
https://orcid.org/0000-0002-8902-6965
Jessica R. Castle
https://orcid.org/0000-0003-1179-5374
Francis J. Doyle
https://orcid.org/0000-0002-3293-9114
Michael C. Riddell
https://orcid.org/0000-0001-6556-7559
Supplemental Material: Supplemental material for this article is available online.
References
- 1. Finn M, Sherlock M, Feehan S, Guinan EM, Moore KB. Adherence to physical activity recommendations and barriers to physical activity participation among adults with type 1 diabetes. Ir J Med Sci. 2022;191(4):1639-1646. doi: 10.1007/s11845-021-02741-w. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Lascar N, Kennedy A, Hancock B, et al. Attitudes and barriers to exercise in adults with type 1 diabetes (T1DM) and how best to address them: a qualitative study. PLoS ONE. 2014;9(9):e108019. doi: 10.1371/journal.pone.0108019. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Tonoli C, Heyman E, Roelands B, et al. Effects of different types of acute and chronic (training) exercise on glycaemic control in type 1 diabetes mellitus: a meta-analysis. Sports Med. 2012;42(12):1059-1080. doi: 10.1007/bf03262312. [DOI] [PubMed] [Google Scholar]
- 4. García-García F, Kumareswaran K, Hovorka R, Hernando ME. Quantifying the acute changes in glucose with exercise in type 1 diabetes: a systematic review and meta-analysis. Sports Med. 2015;45(4):587-599. doi: 10.1007/s40279-015-0302-2. [DOI] [PubMed] [Google Scholar]
- 5. Eckstein ML, Weilguni B, Tauschmann M, et al. Time in range for closed-loop systems versus standard of care during physical exercise in people with type 1 diabetes: a systematic review and meta-analysis. J Clin Med. 2021;10(11):2445. doi: 10.3390/jcm10112445. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Davey RJ, Paramalingam N, Retterath AJ, et al. Antecedent hypoglycaemia does not diminish the glycaemia-increasing effect and glucoregulatory responses of a 10 s sprint in people with type 1 diabetes. Diabetologia. 2014;57(6):1111-1118. doi: 10.1007/s00125-014-3218-8. [DOI] [PubMed] [Google Scholar]
- 7. Sigal RJ, Purdon C, Fisher SJ, Halter JB, Vranic M, Marliss EB. Hyperinsulinemia prevents prolonged hyperglycemia after intense exercise in insulin-dependent diabetic subjects. J Clin Endocrinol Metab. 1994;79(4):1049-1057. doi: 10.1210/jcem.79.4.7962273. [DOI] [PubMed] [Google Scholar]
- 8. Turner D, Gray BJ, Luzio S, et al. Similar magnitude of post-exercise hyperglycemia despite manipulating resistance exercise intensity in type 1 diabetes individuals. Scand J Med Sci Sports. 2016;26(4):404-412. doi: 10.1111/sms.12472. [DOI] [PubMed] [Google Scholar]
- 9. Yardley JE, Kenny GP, Perkins BA, 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. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Yardley JE. Reassessing the evidence: prandial state dictates glycaemic responses to exercise in individuals with type 1 diabetes to a greater extent than intensity. Diabetologia. 2022;65(12):1994-1999. doi: 10.1007/s00125-022-05781-8. [DOI] [PubMed] [Google Scholar]
- 11. Harmer AR, Chisholm DJ, McKenna MJ, et al. High-intensity training improves plasma glucose and acid-base regulation during intermittent maximal exercise in type 1 diabetes. Diabetes Care. 2007;30(5):1269-1271. doi: 10.2337/dc06-1790. [DOI] [PubMed] [Google Scholar]
- 12. Riddell MC, Pooni R, Yavelberg L, et al. Reproducibility in the cardiometabolic responses to high-intensity interval exercise in adults with type 1 diabetes. Diabetes Res Clin Pract. 2019;148:137-143. doi: 10.1016/j.diabres.2019.01.003. [DOI] [PubMed] [Google Scholar]
- 13. Guelfi KJ, Jones TW, Fournier PA. The decline in blood glucose levels is less with intermittent high-intensity compared with moderate exercise in individuals with type 1 diabetes. Diabetes Care. 2005;28(6):1289-1294. doi: 10.2337/diacare.28.6.1289. [DOI] [PubMed] [Google Scholar]
- 14. McClure RD, Alcántara-Cordero FJ, Weseen E, et al. Systematic review and meta-analysis of blood glucose response to high-intensity interval exercise in adults with type 1 diabetes. Can J Diabetes. 2023;47(2):171-179. doi: 10.1016/j.jcjd.2022.11.006. [DOI] [PubMed] [Google Scholar]
- 15. Moser O, Tschakert G, Mueller A, et al. Effects of high-intensity interval exercise versus moderate continuous exercise on glucose homeostasis and hormone response in patients with type 1 diabetes mellitus using novel ultra-long-acting insulin. PLoS ONE. 2015;10(8):e0136489. doi: 10.1371/journal.pone.0136489. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. MacDonald MJ. Postexercise late-onset hypoglycemia in insulin-dependent diabetic patients. Diabetes Care. 1987;10(5):584-588. doi: 10.2337/diacare.10.5.584. [DOI] [PubMed] [Google Scholar]
- 17. Molveau J, Rabasa-Lhoret R, Myette-Côté É, et al. Prevalence of nocturnal hypoglycemia in free-living conditions in adults with type 1 diabetes: what is the impact of daily physical activity. Front Endocrinol (Lausanne). 2022;13:953879. doi: 10.3389/fendo.2022.953879. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Cade WT, Khoury N, Nelson S, et al. Hypoglycemia during moderate intensity exercise reduces counterregulatory responses to subsequent hypoglycemia. Physiol Rep. 2016;4(17):e12848. doi: 10.14814/phy2.12848. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Younk LM, Mikeladze M, Tate D, Davis SN. Exercise-related hypoglycemia in diabetes mellitus. Expert Rev Endocrinol Metab. 2011;6(1):93-108. doi: 10.1586/eem.10.78. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Riddell MC, Li Z, Gal RL, et al. Examining the acute glycemic effects of different types of structured exercise sessions in type 1 diabetes in a real-world setting: the Type 1 Diabetes and Exercise Initiative (T1DEXI). Diabetes Care. 2023;46(1):704-713. doi: 10.2337/dc22-1721. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Abraham MB, Davey RJ, Cooper MN, et al. Reproducibility of the plasma glucose response to moderate-intensity exercise in adolescents with Type 1 diabetes. Diabet Med. 2017;34(9):1291-1295. doi: 10.1111/dme.13395. [DOI] [PubMed] [Google Scholar]
- 22. Soon WHK, Fournier PA, Abraham MB, et al. Reproducibility of plasma glucose responses to moderate-intensity exercise in individuals with type 1 diabetes. Diabet Med. 2023;40(3):e15000. doi: 10.1111/dme.15000. [DOI] [PubMed] [Google Scholar]
- 23. Temple MY, Bar-Or O, Riddell MC. The reliability and repeatability of the blood glucose response to prolonged exercise in adolescent boys with IDDM. Diabetes Care. 1995;18(3):326-332. doi: 10.2337/diacare.18.3.326. [DOI] [PubMed] [Google Scholar]
- 24. Tyler NS, Mosquera-Lopez C, Young GM, El Youssef J, Castle JR, Jacobs PG. Quantifying the impact of physical activity on future glucose trends using machine learning. iScience. 2022;25(3):103888. doi: 10.1016/j.isci.2022.103888. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Biankin SA, Jenkins AB, Campbell LV, Choi KL, Forrest QG, Chisholm DJ. Target-seeking behavior of plasma glucose with exercise in type 1 diabetes. Diabetes Care. 2003;26(2):297-301. doi: 10.2337/diacare.26.2.297. [DOI] [PubMed] [Google Scholar]
- 26. Notkin GT, Kristensen PL, Pedersen-Bjergaard U, Jensen AK, Molsted S. Reproducibility of glucose fluctuations induced by moderate intensity cycling exercise in persons with type 1 diabetes. J Diabetes Res. 2021;2021:6640600. doi: 10.1155/2021/6640600. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27. Gillingham MB, Li Z, Beck RW, et al. Assessing mealtime macronutrient content: patient perceptions versus expert analyses via a novel phone app. Diabetes Technol Ther. 2021;23(2):85-94. doi: 10.1089/dia.2020.0357. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28. Walsh J, Roberts R, Heinemann L. Confusion regarding duration of insulin action: a potential source for major insulin dose errors by bolus calculators. J Diabetes Sci Technol. 2014;8(1):170-178. doi: 10.1177/1932296813514319. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29. Benjamini Y, Hochberg Y. On the adaptive control of the false discovery rate in multiple testing with independent statistics. J Educ Behav Stat. 2000;25(1):60-83. [Google Scholar]
- 30. Prevost MS, Rabasa-Lhoret R, Talbo MK, Yardley JE, Curry EG, Brazeau AS. Gender differences in strategies to prevent physical activity-related hypoglycemia in patients with type 1 diabetes: a BETTER study. Diabetes Care. 2022;45(3):e51-e53. doi: 10.2337/dc21-1899. [DOI] [PubMed] [Google Scholar]
- 31. Frank S, Jbaily A, Hinshaw L, Basu R, Basu A, Szeri AJ. Modeling the acute effects of exercise on insulin kinetics in type 1 diabetes. J Pharmacokinet Pharmacodyn. 2018;45(6):829-845. doi: 10.1007/s10928-018-9611-z. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32. Davey RJ, Howe W, Paramalingam N, et al. The effect of midday moderate-intensity exercise on postexercise hypoglycemia risk in individuals with type 1 diabetes. J Clin Endocrinol Metab. 2013;98(7):2908-2914. doi: 10.1210/jc.2013-1169. [DOI] [PubMed] [Google Scholar]
- 33. McMahon SK, Ferreira LD, Ratnam N, et al. Glucose requirements to maintain euglycemia after moderate-intensity afternoon exercise in adolescents with type 1 diabetes are increased in a biphasic manner. J Clin Endocrinol Metab. 2007;92(3):963-968. doi: 10.1210/jc.2006-2263. [DOI] [PubMed] [Google Scholar]
- 34. Scott SN, Anderson L, Morton JP, Wagenmakers AJM, Riddell MC. Carbohydrate restriction in type 1 diabetes: a realistic therapy for improved glycaemic control and athletic performance? Nutrients. 2019;11(5):1022. doi: 10.3390/nu11051022. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35. Mattsson S, Jendle J, Adolfsson P. Carbohydrate loading followed by high carbohydrate intake during prolonged physical exercise and its impact on glucose control in individuals with diabetes type 1: an exploratory study. Front Endocrinol (Lausanne). 2019;10:571. doi: 10.3389/fendo.2019.00571. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Supplemental material, sj-docx-1-dst-10.1177_19322968241234687 for Factors Affecting Reproducibility of Change in Glucose During Exercise: Results From the Type 1 Diabetes and EXercise Initiative by Zoey Li, Peter Calhoun, Michael R. Rickels, Robin L. Gal, Roy W. Beck, Peter G. Jacobs, Mark A. Clements, Susana R. Patton, Jessica R. Castle, Corby K. Martin, Melanie B. Gillingham, Francis J. Doyle and Michael C. Riddell in Journal of Diabetes Science and Technology