Journal Logo


Pancreatic β-Cell Function Is Associated with Augmented Counterregulation to In-Exercise Hypoglycemia in Type 1 Diabetes


Author Information
Medicine & Science in Sports & Exercise: July 2021 - Volume 53 - Issue 7 - p 1326-1333
doi: 10.1249/MSS.0000000000002613
  • Free


Although conventionally considered a condition of absolute insulin deficiency, it is now recognized that a large portion of individuals with long-standing type 1 diabetes (T1D) retain signs of pancreatic β-cell activity (1). Furthermore, the extent of residual β-cell function, as quantified via plasma c-peptide levels, is positively correlated with glycemic stability as well as the avoidance of (2,3), and/or correction from, hypoglycemia (2,4–8). However, less work has assessed these correlations when glycemia is changing both rapidly and dynamically, as is the case during physical exercise. Although a beneficial effect of regular exercise engagement is successive reductions in blood glucose (BG) that aid the attainment of target HbA1c (9), the heightened metabolic demands of exercising skeletal muscle tissue can potentiate increases in intramuscular glucose uptake through mechanisms independent of, but additive to, insulin (10–12). To avoid a rate of glucose disappearance (Rd) greater than the rate of its appearance (Ra), the body initiates a profound regulatory effect on several cardiovascular and neuroendocrine processes (13). However, because of an inability to autoregulate exogenous insulin concentrations, overt hyperinsulinemia and, therefore, a suppression in Ra during exercise is common in T1D. (14,15) Beyond these acutely apparent concerns, the legacy effects of exercise in enhancing tissue sensitivity to insulin may persist for many hours subsequent to its cessation. When combined with the acknowledgment that antecedent exercise can impair subsequent glucose counterregulation (16,17), there is an elevated risk of postexercise and nocturnal hypoglycemia in this cohort (18,19). With these mechanisms in mind, retention of β-cell function and therefore, some maintenance of intrinsic glucose regulation, may serve to benefit glycemia around physical exercise.

Indeed, recent work has documented improved interstitial glucose (iG) variability after acute aerobic exercise in those with higher endogenous insulin secretory capacity (20). However, the impact of residual pancreatic function on the counterregulatory hormonal responses to hypoglycemia both within and after physical exercise is yet to be established.

Thus, this study sought to investigate the influence of residual β-cell function on counterregulatory hormonal responses to hypoglycemia during acute physical exercise in people with T1D. A secondary aim was to explore relationships between biomarkers of pancreatic β-cell function and indices of glycemia following acute exercise including the nocturnal period.


Study design

This study was a secondary analysis of data collated from a single-center, randomized, open-label, four-period crossover clinical trial consisting of 23-h inpatient phases with a 45-min bout of evening exercise. The primary outcome was to detail the extent and prevalence of postexercise and nocturnal hypoglycemia after periexercise bolus insulin dose adjustments in individuals with T1D using multiple daily injections of insulin aspart (IAsp) and insulin degludec (IDeg). Data from the four interventional trial arms were pooled for this secondary analysis. The study was carried out in accordance with the Declaration of Helsinki and the International Conference on Harmonization Good Clinical Practice (ICH GCP E6 guidelines). Approval was granted by both the national research ethics committee (16/WA/0394) and the local health authority (EudraCT number, 2017-004774-34; UTN, U1111-1174-6676), then registered at the German Clinical Trials Register (; DRKS00013509). All participants were provided with a full written and verbal description of the study and gave informed consent before taking part.

Study inclusion and exclusion criteria

Ahead of trial inclusion, participants were screened for anthropometric, cardiovascular, and T1D specific markers before the performance of a cardiopulmonary exercise test on a semirecumbent cycle ergometer (Corival Recumbent, Lode, NL) to determine peak performance characteristics (21). The main inclusion criteria were as follows: diagnosis of T1D for ≥12 months, age 18 to 65 yr (both inclusive), body mass index of 18.0 to 29.4 kg·m−2, use of multiple daily injections of insulin for ≥12 months, body mass–specific peak oxygen uptake of ≥20 mL·kg−1⋅min−1, and a status of being physically active as assessed by the International Physical Activity Questionnaire Short Form. The main exclusion criteria were as follows: presence of a life-threatening disease, proliferative retinopathy or maculopathy, severe neuropathy, recurrent severe hypoglycemia (more than one severe hypoglycemia event during the previous 12 months), hypoglycemia unawareness as judged by the investigator, hospitalization for diabetic ketoacidosis during the previous 6 months, and any other condition that would interfere with trial participation or evaluation of results as judged by the investigator.

After successful completion against the reference inclusion criteria, participants were switched from their usual basal/bolus insulin therapies to IDeg (Tresiba®; Novo Nordisk, Bagsvaerd, Denmark) and IAsp (NovoRapid®, Novo Nordisk). Participants were also provided with an interstitial glucose monitoring (iGM) device (Freestyle Libre; Abbott Laboratories, Chicago, IL) for the duration of their study involvement.

Trial design

After preliminary testing (visit 1) and a run-in period for adjustment to IDeg, visits 2, 3, 4, and 5 were experimental visits that involved 23 h of inpatient monitoring with an overnight stay in a clinical research facility. After a daytime control period (0800–1559 h) during which all dietary and insulin timings and amounts were standardized, participants undertook a bout of evening (1700 h) cycling exercise at ~60% ± 6% V˙O2max. One hour before and after exercise (Ex), participants administered either a full (100%) or a reduced (50%) dose (100%; 5.1% ± 2.4% vs 50%; 2.6 ± 1.2 IU, P < 0.001) of individualized IAsp alongside the consumption of identical low glycemic index (brown rice-based vegetable dish), carbohydrate-rich meals equating to 1 g CHO·kg−1⋅bm−1 (total energy, 496 ± 62 kcal; fat, 9 ± 5 g [20%]; protein, 19 ± 11 g [15%]; CHO, 80 ± 10 g [65%]). An unaltered and regular dose of IDeg was kept consistent across each experimental visit. Trial-day glycemia was determined via capillary (0800–1559 h), venous (1600–0700 h), and interstitial (0800–0700 h) glucose monitoring. Hypoglycemia was defined as a venous BG value of ≤3.9 mmol⋅L−1 (22).

Participant characteristics and C-peptide stratification

As part of this exploratory, secondary analysis, retrospectively identified samples taken from the acute exercise period were used to compare biomarker responses between individuals with T1D when stratified based on their residual pancreatic β-cell function. Participants were split into two groups based on their stimulated (1-h postprandial) C-peptide values. Individuals with a mean value of ≥30.0 pmol⋅L−1 were considered low-level secretors (LLS), whereas those with a value of <30.0 pmol⋅L−1 were classified as microsecretors (MS) (1). Baseline characteristics of study participants when split based on their stimulated C-peptide status are displayed in Table 1. Diabetes duration was significantly shorter in LLS (P = 0.01, Table 1)

TABLE 1 - Baseline characteristics of study participants when split based on their stimulated C-peptide status.
Characteristic LLS MS P
Number, n (M/F) 6 (5/1) 10 (8/2)
Age (yr) 30.3 ± 14.3 37.0 ± 13.8 0.38
BMI (kg·m−2) 25.6 ± 4.2 26.2 ± 3.1 0.77
Lean mass (%) 23.3 ± 3.6 23.5 ± 3.3 0.91
HbA1c (%) 6.9 ± 1.7 7.5 ± 1.1 0.51
Diabetes duration (yr) 7.0 ± 4.6 18.9 ± 11.6 0.01
Age of onset (yr) 23.3 ± 16.0 18.1 ± 6.7 0.48
Prestudy TDD (IU·kg−1⋅bm−1) 39 ± 18 59 ± 29 0.11
Fasted BG (mmol⋅L−1) 8.9 ± 3.0 9.2 ± 2.9 0.87
Fasted C-peptide (pmol⋅L−1) 173.1 ± 128.4 11.2 ± 7.3 0.03
Fasted proinsulin (pmol⋅L−1) 4.3 ± 6.2 0.03 ± 0.03 0.05
V˙O2max (mL·kg−1⋅min−1) 45.7 ± 12.4 37.1 ± 7.9 0.17
Powermax (W·kg−1⋅bm−1) 3.3 ± 0.9 2.7 ± 0.6 0.22
Data are presented as mean ± SD.
M, males; F, females; BMI, body mass index; HbA1c, glycated hemoglobin; TDD, total daily dose of insulin (inclusive of basal and bolus amounts); IU, insulin units; V˙O2max, maximum oxygen consumption; P-amy, pancreatic α-amylase.

Moderate-intensity continuous exercise testing protocol

The exercise test consisted of 45 min (3-min warm-up at 20 W, 42 min at target workload) of continuous cycling on a semirecumbent ergometer (Corival Recumbent) at 60% ± 6% V˙O2max. The workload intensity was computed as the midpoint between the first and the second lactate turn points as previously described (21). Participants were instructed to maintain a stable pedal cadence between 65 and 75 rpm against a fixed workload. During exercise, fingertip capillary blood glucose sampling was performed every 6 min and measured instantaneously via the inbuilt glucometer. If capillary BG reached the hypoglycemic threshold (≤3.9 mmol⋅L−1) before completion of the full test protocol, exercise testing was terminated immediately. At this point, a venous blood sample was drawn instantly. In the event of hypoglycemia, participants received a standardized CHO gel as a rescue aid (Glucogel®; BBI Healthcare Ltd., Crumlin, UK) and were monitored to ensure glycemic recovery.

Venous-derived biomarker analysis

Before laboratory analysis, retrospectively identified frozen plasma samples were thawed at room temperature, centrifuged for 2 min at 4000 rpm, and then treated as per the individualized assay kit instructions. The Randox Daytona Plus RX series analyzer (Randox Laboratories Ltd.®, Crumlin, UK) was used for in vitro determination of β-hydroxybutyrate ([β-OHB] RB4067). C-peptide was quantified using Invitron’s Molecular Light Technology™ chemiluminescence assay (IV2-004; Invitron Ltd.®, Monmouth, UK). Proinsulin was quantified using Invitron’s Molecular Light Technology™ chemiluminescence intact assay (IV2-002, Invitron Ltd.). Glucagon was quantified using the R&D Systems® (R&D Systems, Inc., Minneapolis, MN) Glucagon Quantikine ELISA Kit (product no. DGCG0). The Eagle Biosciences (Eagle Biosciences, Inc., Amherst, NH) BI-CAT Epinephrine & Norepinephrine ELISA Assay Kit (product no. ECT31-K02) was used for the determination of catecholamines. Trial-day venous BG and blood lactate (BLa) concentrations were measured via the fully enzymatic–amperometric method (Biosen C-Line; EKF-diagnostic GmbH, Barleben, Germany).

Statistical analysis

All statistical analyses were carried out using IBM SPSS 26.0 (IBM Corp., Armonk, NY) and a P value of ≤0.05 was accepted for statistical significance. A repeated-measures ANOVA on four levels was conducted, with Bonferroni-corrected pairwise comparisons used to examine time effects. Baseline characteristics data and iGM metrics were assessed via independent t-tests. Trial-day venous-derived biomarker data were pooled from the four experimental trial arms, hence treated via uni- or multivariate analysis with the preexercise bolus insulin dose included as a covariate in the model for all postbaseline sample time points. Pearson’s correlation coefficients were used to explore the relationship between variables at baseline and in response to exercise. Fisher’s r to z transformation was performed to determine statistical significance in the correlation coefficient between groups using the following formula: Zobserved = (z1z2)/(square root of [(1/N1 – 3) + (1/N2 – 3)] (23). Cross-tabulation analysis was used to identify estimated risk ratios (ERR) between nominal variables, with Fisher’s exact testing used to report significance. iGM-derived data taken from the 23 h during experimental trial days were used to assess indices of glycemic control. iG data were stratified into predefined glycemic ranges (24), i.e., time below target glucose ([TBR] <3.9 mmol⋅L−1), time within target glucose ([TIR] ≥3.9 to ≤10.0 mmol⋅L−1), and time above target glucose ([TAR] ≥10.1 mmol⋅L−1). Trial-day iG data were also stratified into distinct periods, i.e., overall 23-h data (0800–0700 h), in-exercise data (1700–1745 h), postexercise data (1746–2359 h), and nocturnal data (0000–0700 h).


Periexercise biomarker responses

Figure 1 displays the periexercise-induced changes in biomarker concentrations in MS and LLS groups.

Trial-day concentrations of metabolic pancreatic and catecholamine biomarkers when participants were split based on their stimulated C-peptide status. BG (A), β-OHB (B), BLa (C), C-peptide (D), proinsulin (E), glucagon (F), EPI (G), NE (H). Ex, exercise. Data are presented as mean ± SEM. *P ≤ 0.05 in the point concentration between groups. **P ≤ 0.001 in the point concentration between groups. ***P < 0.001 in the point concentration between groups.

At all time points, LLS group had higher C-peptide (P < 0.01), proinsulin (P < 0.01), and BLa (P ≤ 0.05) concentrations compared with MS. Glucagon was higher in LLS compared with MS at baseline (P = 0.02) and preexercise (P = 0.05), with a trend acutely postexercise (P = 0.06). Immediately after exercise, LLS had lower BG (LLS 3.7 ± 1.0 vs MS 4.9 ± 2.6 mmol⋅L−1, P = 0.05) but higher epinephrine (EPI) (LLS 0.12 ± 0.14 vs MS 0.05 ± 0.07 nmol⋅L−1, P = 0.02), β-OHB (LLS 0.05 ± 0.02 vs MS 0.04 ± 0.01 mmol⋅L−1, P < 0.01), and BLa (LLS 3.1 ± 1.8 vs MS 2.4 ± 0.8 mmol⋅L−1, P = 0.05) concentrations. There were no differences in norepinephrine (NE) concentrations between groups.

Exercise responses

LLS had a lower bolus insulin dose than MS with the preexercise meal (LLS, 3.0 ± 1.7, vs MS, 4.4 ± 2.5 IU; P < 0.01), but the preexercise BG concentrations (LLS, 8.2 ± 2.3, vs MS, 8.2 ± 2.9 mmol⋅L−1; P = 0.96; Fig. 1A) and the exercise duration (LLS, 38.8 ± 8.2, vs MS, 38.5 ± 8.2 min; P = 0.90) were identical between groups. Table 2 outlines the change in biomarker in response to exercise (ExΔ).

TABLE 2 - The exercise-induced change in pancreatic, metabolic, and catecholamine concentrations when data have been stratified based on microsecretory capacity.
Biomarker ΔLLS ΔMS P
Pancreatic responses
 C-peptide (pmol⋅L−1) −60.9 ± 50.5 −0.6 ± 8.8 <0.001
 Proinsulin (pmol⋅L−1) −1.5 ± 2.6 −0.1 ± 1.1 0.01
 Glucagon (pg⋅mL−1) −1.5 ± 57.3 +2.8 ± 18.1 0.87
Metabolic responses
 BG (mmol⋅L−1) −4.4 ± 2.2 −3.3 ± 2.2 0.07
 BLa (mmol⋅L−1) +2.0 ± 1.7 +1.5 ± 0.8 0.13
 β-OHB (mmol⋅L−1) +0.014 ± 0.02 +0.003 ± 0.01 0.02
Catecholamine responses
 NE (nmol⋅L−1) +1.0 ± 1.3 +0.3 ± 1.2 0.11
 EPI (nmol⋅L−1) +0.1 ± 0.1 −0.0 ± 0.1 0.01
Data are presented as mean ± SD.
Bold italics represent a significant difference (P ≤ 0.05) in data values between the LLS and MS groups.

Exercise-induced increases in plasma catecholamines (EPI: LLS = +181%, P = 0.037, MS = −3%, P = 1.00; NE: LLS = +155%, P = 0.001, MS = +38%, P = 0.46) and β-OHB (LLS = +36%, P = 0.015; MS = +9%, P = 1.00) occurred in LLS only. In both groups, BLa rose (LLS = +186, MS = +177; both P < 0.001), whereas BG dropped (LLS = −54%, MS = −40%; both P < 0.001) during exercise. Glucagon remained unchanged in both groups (P ≥ 0.05).

LLS had a greater suppressive effect on C-peptide and proinsulin, as well as significantly greater EPI and β-OHB responses to exercise (Table 2). The exercise-induced change (ExΔ) in BG, BLa, glucagon, and NE were comparable between groups, as was the rate of change in BG over exercise (LLS = −0.12 ± 0.06 mmol⋅L−1⋅min−1 vs MS = −0.09 ± 0.06 mmol⋅L−1⋅min−1, P = 0.08).

The ExΔ in glucagon was negatively related to the ExΔ in C-peptide (r = −0.320, P = 0.014). Further analysis revealed that the correlation was solely driven by the LLS group (r = −0.544, P = 0.01), whereas no relationship was evident in the MS group (r = −0.030, P = 0.86). The strength of correlation differed between groups (Zobserved = 2.01, P = 0.04). The ExΔ values in EPI (r = −0.361, P = 0.003) and NE (r = −0.360, P = 0.003) were both inversely correlated with the ExΔ in C-peptide.

Influence of residual β-cell function on the risk and rescue response to hypoglycemia

Throughout the experimental period, there were a total of 66 hypoglycemic events (LLS = 21 vs MS = 45 events, P = 0.43). Every participant experienced at least one hypoglycemia event throughout their study involvement. In LLS, 4/6 (67%) individuals experienced recurrent (>1 event) hypoglycemia whereas 8/10 (80%) MS experienced a repeat event. The risk of recurrent hypoglycemia was similar between groups (ERR = 1.1, 95% CI = 0.292–4.151, P = 0.89).

Out of 64 exercise tests, 39 ended prematurely because of hypoglycemia (LLS = 17 vs MS = 22 events, P = 0.30). During the ~13-h postexercise period (including the nocturnal and fasted morning hours), half of the participants (8/16) had a second hypoglycemic event after experiencing antecedent exercise-induced hypoglycemia. Of these 8 participants, 2 (25%) were LLS whereas 6 (75%) were MS. The risk of experiencing subsequent hypoglycemia after antecedent exercise-induced hypoglycemia was similar between LLS and MS groups (ERR = 2.0, 95% CI = 0.500–8.00, P = 0.61). Of the 27 hypoglycemic events that occurred at rest, 4 (15%) occurred in LLS whereas 23 (85%) occurred in MS (P = 0.07); 50% of LLS and 90% of the MS experienced a rested hypoglycemic event (ERR = 3.0, 95% CI = 0.968–9.302, P = 0.07). Table 3 details biomarker differences between groups (a) in response to exercise, (b) at rest, and (c) all hypoglycemic events over the entire experimental period.

TABLE 3 - Point concentrations in venous-derived biomarkers during (A) the 39 hypoglycemic events, which occurred during exercise only; (B) the 27 events that occurred at rest, having extracted those which occurred during exercise; and (C) all 66 hypoglycemic events that occurred during the 14-h experimental trial period.
Biomarker LLS MS P
(A) In-exercise hypoglycemia
 BG (mmol⋅L−1) 3.2 ± 0.4 3.4 ± 0.4 0.25
 C-peptide (pmol⋅L−1) 144.7 ± 121.2 11.6 ± 12.1 <0.01
 Proinsulin (pmol⋅L−1) 3.3 ± 4.6 0.1 ± 0.3 <0.01
 Glucagon (pg⋅mL−1) 39.6 ± 63.7 9.7 ± 17.8 0.09
 NE (nmol⋅L−1) 1.6 ± 1.4 0.9 ± 0.7 0.05
 EPI (nmol⋅L−1) 0.12 ± 0.14 0.05 ± 0.07 0.02
 β-OHB (mmol⋅L−1) 0.05 ± 0.05 0.01 ± 0.03 <0.01
 BLa (mmol⋅L−1) 2.9 ± 1.7 2.6 ± 0.7 0.39
(B) Rested (nonexercise) hypoglycemia
 BG (mmol⋅L−1) 2.9 ± 0.2 3.3 ± 0.4 0.10
 C-peptide (pmol⋅L−1) 89.8 ± 132.3 14.5 ± 8.9 0.01
 Proinsulin (pmol⋅L−1) 3.60 ± 6.24 0.04 ± 0.04 0.01
 Glucagon (pg⋅mL−1) 61.6 ± 120.4 40.3 ± 88.4 0.68
 NE (nmol⋅L−1) 0.78 ± 0.68 0.41 ± 0.51 0.21
 EPI (nmol⋅L−1) 0.07 ± 0.08 0.04 ± 0.05 0.38
 β-OHB (mmol⋅L−1) 0.04 ± 0.00 0.05 ± 0.02 0.56
 BLa (mmol⋅L−1) 0.92 ± 0.45 0.61 ± 0.28 0.07
(C) All hypoglycemic events
 BG (mmol⋅L−1) 3.1 ± 0.4 3.3 ± 0.4 0.07
 C-peptide (pmol⋅L−1) 134.2 ± 122.0 13.1 ± 10.6 <0.01
 Proinsulin (pmol⋅L−1) 3.3 ± 4.8 0.1 ± 0.2 <0.01
 Glucagon (pg⋅mL−1) 43.7 ± 74.1 25.7 ± 66.3 0.33
 NE (nmol⋅L−1) 1.4 ± 1.3 0.6 ± 0.6 <0.01
 EPI (nmol⋅L−1) 0.11 ± 0.13 0.05 ± 0.06 0.01
 β-OHB (mmol⋅L−1) 0.05 ± 0.05 0.03 ± 0.03 0.05
 BLa (mmol⋅L−1) 2.5 ± 1.7 1.6 ± 1.2 0.02
Data are presented as mean ± SD.
Bold italics represent a significant difference (P ≤ 0.05) in data values between the LLS and MS groups.

During in-exercise hypoglycemia, LLS presented with higher sympathoadrenal (EPI and NE), ketone, and pancreatic β-cell biomarker concentrations than MS, with a trend toward a difference in the glucagon response (Table 3A). Of note, removal of the preexercise IAsp dose as a covariate in the model produced borderline significance in the glucagon concentrations between groups during in-exercise hypoglycemia (P = 0.05).

The responsiveness of EPI (r = −0.504, P < 0.001), NE (r = −0.570, P < 0.001), and glucagon (r = −0.504, P = 0.001) to in-exercise hypoglycemia were inversely correlated to the ExΔ in C-peptide during exercise. Linear regression analysis demonstrated that diabetes duration had no impact of the counterregulatory hormonal responses to hypoglycemia at rest, during exercise, or overall (all P ≥ 0.05).

Influence of residual B-cell function on interstitial-derived indices of glycemic control

Table 4 details iGM-derived indices of glycemic control taken at distinct time phase during the 23-h experimental trial period. LLS has significantly less glycemic instability in iG over 23 h of inpatient monitoring, including lower peaks and higher minimum glucose values. There were no differences between groups in iG metrics during the exercise period. LLS spent less TBR in the acute postexercise period and had lower glycemic variability (CoV and SD) in iG data throughout the nocturnal hours.

TABLE 4 - Time spent in each glycemic range during the 23-h experimental trial period.
Parameter LLS MS P
Exercise (1700–1745 h)
Exercise mean iG (mmol⋅L−1) 9.8 ± 2.1 9.6 ± 3.1 0.80
Exercise SD (mmol⋅L−1) 1.9 ± 0.8 1.6 ± 0.8 0.19
Exercise CoV (%) 18.9 ± 6.6 17.3 ± 8.5 0.44
Exercise minimum (mmol⋅L−1) 6.7 ± 1.8 7.1 ± 3.2 0.61
Exercise maximum (mmol⋅L−1) 11.6 ± 2.5 11.2 ± 3.2 0.60
Exercise TBR (%) 0.0 ± 0.0 0.3 ± 2.1 0.61
Exercise TIR (%) 53.9 ± 33.2 58.5 ± 34.4 0.43
Exercise TAR (%) 46.1 ± 33.2 42.5 ± 33.8 0.70
Postexercise (1746–2359 h)
 Post-Ex mean iG (mmol⋅L−1) 10.1 ± 2.0 10.2 ± 3.9 0.88
 Post-Ex SD (mmol·L−1) 2.2 ± 0.8 2.3 ± 1.0 0.54
 Post-Ex CoV (%) 21.5 ± 6.7 23.5 ± 8.1 0.30
 Post-Ex minimum (mmol⋅L−1) 5.7 ± 1.4 5.6 ± 2.9 0.92
 Post-Ex maximum (mmol⋅L−1) 13.5 ± 2.4 13.6 ± 4.3 0.96
 Post-Ex TBR (%) 0.2 ± 0.8 3.2 ± 6.3 0.01
 Post-Ex TIR (%) 51.1 ± 31.3 53.3 ± 36,9 0.80
 Post-Ex TAR (%) 48.7 ± 31.6 43.5 ± 39.5 0.57
Night (0000–0700 h)
 Night mean iG (mmol⋅L−1) 11.4 ± 3.1 10.2 ± 3.7 0.18
 Night SD (mmol⋅L−1) 1.1 ± 0.4 1.8 ± 0.9 <0.001
 Night CoV (%) 10.8 ± 5.4 19.9 ± 11.9 <0.001
 Night minimum (mmol⋅L−1) 9.6 ± 3.2 7.5 ± 3.9 0.03
 Night maximum (mmol⋅L−1) 13.5 ± 3.1 13.5 ± 3.6 0.96
 Night TBR (%) 2.2 ± 10.9 2.9 ± 7.7 0.79
 Night TIR (%) 29.1 ± 35.5 47.6 ± 40.5 0.07
 Night TAR (%) 68.7 ± 38.2 49.5 ± 42.7 0.08
23 h (0800–0700 h)
 Overall mean iG (mmol⋅L−1) 10.2 ± 1.7 9.9 ± 2.2 0.62
 Overall SD (mmol⋅L−1) 2.4 ± 0.7 3.3 ± 0.9 <0.001
 Overall CoV (%) 24.7 ± 4.8 33.6 ± 7.5 <0.001
 Overall minimum (mmol⋅L−1) 4.8 ± 0.8 4.1 ± 1.2 0.01
 Overall maximum (mmol⋅L−1) 14.8 ± 2.1 16.7 ± 3.2 0.01
 Overall TBR (%) 0.9 ± 3.8 2.1 ± 3.5 0.20
 Overall TIR (%) 48.6 ± 21.1 54.9 ± 21.7 0.27
 Overall TAR (%) 50.5 ± 22.9 43.0 ± 23.1 0.22
Bold italics represent a significant difference (P ≤ 0.05) in data values between the LLS and MS groups.
Data have also been stratified into distinct periods. Data are reported as mean ± SD; n = 16. CoV, coefficient of variation; TBR, time below range level 1 (<3.9 mmol⋅L−1); TIR, time in range (≥3.9 to ≤10.0 mmol⋅L−1); TAR, time above range level 1 (≥10.1 mmol⋅L−1).


This study sought to explore the influence of residual β-cell function on the counterregulatory responses to hypoglycemia during acute physical exercise in people with T1D. A secondary aim was to explore relationships between of biomarkers of pancreatic β-cell function and indices of glycemia following acute exercise including the nocturnal period.

Our primary finding showed that residual β-cell function was associated with greater sympathoadrenal and ketonic responses to exercise-induced hypoglycemia in people with T1D. In addressing our secondary aim, we found that those with greater endogenous insulin secretory capacity spent significantly less in the interstitial glucose-derived hypoglycemic range over the early (~6 h) postexercise period and had considerably lower glycemic variability throughout the nocturnal hours. Thus, our collective data demonstrate that the degree of β-cell function corresponds favorably with glycemic outcomes both during and after moderate-intensity continuous exercise in people with T1D.

The appreciable differences between groups in the catecholamines, particularly the adrenomedullary responses to both exercise per se and exercise-induced hypoglycemia, suggest that an attenuation in sympathoadrenal system sensitivity may be an acquired defect that develops concomitant with β-cell failure. Indeed, our participants were matched in HbA1c and free from any major diabetic complications, and although the duration of diabetes did differ, our regression analysis revealed this had no apparent effect on the counterregulatory hormonal responses to hypoglycemia. Rather, the greatest determinant of the adrenomedullary responses to hypoglycemia and exercise appeared to be the degree of residual pancreatic β-cell function.

Our exercise model potentiated large (~45%) reductions in BG concentrations in both groups. Somewhat unexpectedly, the nonhypoglycemic end of exercise BG value was significantly lower in those with higher residual pancreatic function. Although difficult to interpret, we postulate that this could be the result of a greater relative total “on-board” insulin during exercise, i.e., the cumulative effect of an increased appearance of exogenous insulin in the circulation due to the hemodynamic alteration induced by physical exercise, combined with higher absolute endogenous pancreatic insulin concentrations as a result of being in the acute postabsorptive phase having just eaten a high carbohydrate meal. We must also recognize the interventional manipulations made to the insulin dose leading into exercise and consider that although outside of statistical significance, the lower BG concentrations in the LLS group may have prompted a larger glucoregulatory response by virtue of a lower glucose concentration. Whilst outside of statistical significance, the trend for slightly lower hypoglycemic blood glucose concentrations in the LLS group may have prompted a larger glucoregulatory response in line with the hierarchical nature of glucose counter-regulation.

Despite our low-level and microsecretory groups having equivalent drops in BG during exercise, neither glucagon nor the catecholamines initiated any form of response to exercise in those with undetectable C-peptide levels, who also presented with a much lower sympathoadrenal response to within-exercise hypoglycemia relative to those with retained β-cell function. Given that the end of exercise glucose concentrations bordered severe hypoglycemia (<3.0 mmol⋅L−1) in both groups, these data may support the concept that the glucose threshold for counterregulation is considerably lowered in those with T1D (25), yet we show that this is most evident in those with progressed β-cell loss.

Of note, the increased appearance of β-OHB upon exercise cessation in the LLS may suggest higher rates of fatty acid oxidation in hepatic mitochondria (the combined product of lower BG, a greater suppression of endogenous insulin secretion, and augmented EPI and glucagon responses) and, hence, an amplification of β-OHB in plasma (26). Although we did not observe any discernible differences between groups in the glucagon responses to either hypoglycemia or exercise, the suppression of C-peptide during exercise was inversely associated with the subsequent responsiveness of glucagon to exercise per se, as well exercise-induced hypoglycemia in those with residual β-cell-function. Yet evidence of this intraislet cross-talk was completely unapparent in those with undetectable C-peptide concentrations. It is proposed that loss of this intraislet communication plausibly explains the lack of glucagon responsiveness to hypoglycemia in those with T1D (27). Thus, these data may suggest that even minimal amounts of endogenous β-cell function and, hence, greater scope for intrapancreatic signaling can aid glucose counterregulation during a metabolically demanding exercising environment.

Interestingly, despite the clear contrast between groups in the glucoregulatory hormonal responses to hypoglycemia during exercise, as well as the proportionate hypoglycemic glucose values between rested and exercise-induced hypoglycemic events, we observed no differences in the low-level and microsecretory groups when hypoglycemia occurred outside of the exercising environment. This might suggest that the glucoregulatory advantage of residual β-cell function may be most apparent when the system is exposed to the cumulative layer of metabolic and physiological stress brought about by dynamic physical exercise, whereas the independent stimulus of hypoglycemia per se may be insufficient to evoke adequate autonomic system activation. Furthermore, recent antecedent hypoglycemia (28), and exercise (16,17), can reduce the autonomic responses to, symptoms of, and physiological defense against, subsequent hypoglycemia in patients with T1D. This may explain the comparability in the glucoregulatory hormonal responses to hypoglycemia beyond the exercise setting in our data set, because irrespective of residual β-cell function, both groups were exposed to an identical exercise stimulus that led to an equivalent occurrence rate of hypoglycemia. Of note, the risk of hypoglycemia outside of the exercise period was threefold greater in those with undetectable residual β-cell function. From a clinical perspective, these data might suggest a greater risk of hypoglycemia in exercising patients with advanced β-cell loss. Although this should not be a deterrent in the advocacy of frequent exercise engagement, it does highlight the need for extra diligence in glucose monitoring and thorough patient education in strategies that help mitigate the risk of exercise related hypoglycemia.

As part of a secondary aim, we explored the relationship of pancreatic function to indices of glycemia in the hours leading into and throughout the nocturnal period after performance of acute evening exercise. Our findings revealed that those with higher residual pancreatic β-cell function spent less time in the interstitial-derived hypoglycemic range within the postexercise (~6 h) period and had significantly lower glycemic variability in iG data over an entire days’ worth of inpatient monitoring, including the designated nocturnal period. Our study design exposed participants to multiple glycemic challenges, including a metabolically demanding exercise bout, several carbohydrate-rich meals, and predefined manipulations to periexercise bolus IAsp doses. With this in mind, even a minor amount endogenous insulin secretion, and hence, a greater degree of intrinsic glucose regulation, may not only help reduce the risk of early onset postexercise hypoglycemia but also offset the severity of nocturnal glycemic excursions after evening exercise in individuals with T1D. These data align with previous works by Brooks and colleagues (29) that documented an intrinsic relationship between endogenous C-peptide concentrations and lower glycemic variability in islet transplant recipients with T1D. In an exercising context, Taylor et al. (20) recently reported that compared with patients with low (10–190 pmol⋅L−1) and undetectable (<10 pmol⋅L−1) C-peptide values, those with the highest levels (≥200 pmol⋅L−1) spent significantly greater TIR, less TAR, and had lower standard deviation in iG in the 12 and 24 h after exercise. In the present study, using a much lower detection threshold of 30 pmol⋅L−1, we have shown that even a very minor (30 pmol⋅L−1) amount, endogenous insulin secretion may not only aid glucose counterregulation but also reduce the risk of early onset postexercise hypoglycemia and offset the severity of nocturnal interstitial-derived glycemic excursions after evening exercise in individuals with T1D.

Strengths, limitations, and future directions

This study involved detailed metabolic, pancreatic, and hormonal profiling of retrospectively collated samples in people with and without evidence of residual pancreatic β-cell function under controlled laboratory conditions. When combined with access to 24 hourly iG values, these data provided insight into the relationships between exercise and glucose counterregulation on the basis of endogenous insulin secretory capacity in those with T1D. However, incorporation of a control group and larger participant numbers with greater participant heterogeneity are important factors to be considered for further research.


Higher residual β-cell function was associated with greater sympathoadrenal and ketonic responses to exercise-induced hypoglycemia as well as improved glycemia leading into and throughout the nocturnal hours. Even a minimal amount of residual β-cell function confers a beneficial effect on glycemic outcomes during and after exercise in people with T1D.

This study was funded by Novo Nordisk A/S as part of an Investigational Sponsored Study. The authors thank the participants for their significant time commitments and adherence to the study protocol. They also thank the health care professional staff from the Joint Clinical Research Facility and the laboratory research team from the Diabetes Research Unit at Swansea University for their invaluable contributions to the obtention and analysis of data, respectively. A final thanks to Dr. Daniel West for his critical review and feedback on our manuscript. The results of the present study do not constitute endorsement by the American College of Sports Medicine.

Olivia McCarthy has received a Zienkiewcz scholarship and travel grants from Novo Nordisk UK. Othmar Moser has received lecture fees from Medtronic; travel grants from Novo Nordisk A/S, Novo Nordisk AT, Novo Nordisk UK, Medtronic AT; research grants from Sêr Cymru II COFUND fellowship/European Union, Novo Nordisk A/S, and Novo Nordisk AT; as well as material funding from Abbott Diabetes Care. Max L. Eckstein has received a KESS2/European Social Fund scholarship and travel grants from Novo Nordisk A/S. Stephen C. Bain has received research grants (includes principal investigator, collaborator, or consultant and pending grants as well as grants already received) from Health care and Research Wales (Welsh Government) and Novo Nordisk. He has received other research support from Healthcare and Research Wales (Welsh Government) and honoraria from Novo Nordisk, Sanofi, Lilly, Boehringer-Ingelheim, and Merck, and he has an ownership interest in Glycosmedia (diabetes online news service). Richard M. Bracken reports having received honoraria, travel, and educational grant support from Boehringer-Ingelheim, Eli Lilly and Company, Novo Nordisk, and Sanofi-Aventis. The remaining authors have no relevant conflict of interest to disclose. The findings of the present study have been presented clearly, honestly, and without fabrication, falsification, or inappropriate data manipulation. The sponsors did not have any involvement in the conduct of the study nor in the interpretation of the results.


1. Oram RA, Jones AG, Besser REJ, et al. The majority of patients with long-duration type 1 diabetes are insulin microsecretors and have functioning beta cells. Diabetologia. 2014;57(1):187–91.
2. Steffes MW, Sibley S, Jackson M, Thomas W. β-cell function and the development of diabetes-related complications in the diabetes control and complications trial. Diabetes Care. 2003;26(3):832–6.
3. Rickels MR, Evans-Molina C, Bahnson HT, et al. High residual C-peptide likely contributes to glycemic control in type 1 diabetes. J Clin Invest. 2020;130(4):1850–62.
4. Fukuda M, Tanaka A, Tahara Y, et al. Correlation between minimal secretory capacity of pancreatic beta-cells and stability of diabetic control. Diabetes. 1988;37(1):81–8.
5. Gonen B, Goldman J, Baldwin D, et al. Metabolic control in diabetic patients. Effect of insulin-secretory reserve (measured by plasma C-peptide levels) and circulating insulin antibodies. Diabetes. 1979;28(8):749–53.
6. Madsbad S, McNair P, Faber OK, Binder C, Christiansen C, Transbøl I. Beta-cell function and metabolic control in insulin treated diabetics. Acta Endocrinol (Copenh). 1980;93(2):196–200.
7. Grajwer LA, Pildes RS, Horwitz DL, Rubenstein AH. Control of juvenile diabetes mellitus and its relationship to endogenous insulin secretion as measured by C-peptide immunoreactivity. J Pediatr. 1977;90(1):42–8.
8. Madsbad S, Hilsted J, Krarup T, et al. Hormonal, metabolic and cardiovascular responses to hypoglycaemia in type 1 (insulin-dependent) diabetes with and without residual B cell function. Diabetologia. 1982;23(6):499–503.
9. Bohn B, Herbst A, Pfeifer M, et al. Impact of physical activity on glycemic control and prevalence of cardiovascular risk factors in adults with type 1 diabetes: a cross-sectional multicenter study of 18,028 patients. Diabetes Care. 2015;38(8):1536–43.
10. Lund S, Holman GD, Schmitz O, Pedersen O. Contraction stimulates translocation of glucose transporter GLUT4 in skeletal muscle through a mechanism distinct from that of insulin. Proc Natl Acad Sci U S A. 1995;92(13):5817–21.
11. Chokkalingam K, Tsintzas K, Norton L, Jewell K, Macdonald IA, Mansell PI. Exercise under hyperinsulinaemic conditions increases whole-body glucose disposal without affecting muscle glycogen utilisation in type 1 diabetes. Diabetologia. 2007;50(2):414–21.
12. DeFronzo RA, Ferrannini E, Sato Y, Felig P, Wahren J. Synergistic interaction between exercise and insulin on peripheral glucose uptake. J Clin Invest. 1981;68(6):1468–74.
13. Hackney AC, Lane AR. Exercise and the regulation of endocrine hormones. In: Progress in Molecular Biology and Translational Science. Elsevier B.V.; 2015. pp. 293–311.
14. Petersen KF, Price TB, Bergeron R. Regulation of net hepatic glycogenolysis and gluconeogenesis during exercise: impact of type 1 diabetes. J Clin Endocrinol Metab. 2004;89(9):4656–64.
15. Camacho RC, Pencek RR, Lacy DB, James FD, Wasserman DH. Suppression of endogenous glucose production by mild hyperinsulinemia during exercise is determined predominantly by portal venous insulin. Diabetes. 2004;53(2):285–93.
16. Sandoval DA, Aftab Guy DL, Richardson MA, Ertl AC, Davis SN. Acute, same-day effects of antecedent exercise on counterregulatory responses to subsequent hypoglycemia in type 1 diabetes mellitus. Am J Physiol Metab. 2006;290(6):E1331–8.
17. Sandoval DA, Aftab Guy DL, Richardson MA, Ertl AC, Davis SN. Effects of low and moderate antecedent exercise on counterregulatory responses to subsequent hypoglycemia in type 1 diabetes. Diabetes. 2004;53(7):1798–806.
18. Campbell MD, Walker M, Trenell MI, et al. Large pre- and postexercise rapid-acting insulin reductions preserve glycemia and prevent early- but not late-onset hypoglycemia in patients with type 1 diabetes. Diabetes Care. 2013;36(8):2217–24.
19. 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):963–8.
20. Taylor GS, Smith K, Scragg J, et al. 294-OR: type 1 diabetes patients with residual beta-cell function display improved time in euglycemia and less glycaemic fluctuation after exercise. Diabetes. 2019;68(1 Suppl):294–OR.
21. Hofmann P, Tschakert G. Special needs to prescribe exercise intensity for scientific studies. Cardiol Res Pract. 2010;2011:209302.
22. Seaquist ER, Anderson J, Childs B, et al. Hypoglycemia and diabetes: a report of a workgroup of the American diabetes association and the endocrine society. Diabetes Care. 2013;36:1384–95.
23. Cohen J, Cohen P, West SG, Aiken LS. In: Riegert D, editor. Applied Multiple Regression/Correlation Analysis for the Behavioral Sciences. 3rd ed. New Jersey: Lawrence Erlbaum Associates; 2003. 240 p.
24. Battelino T, Danne T, Bergenstal RM, et al. Clinical targets for continuous glucose monitoring data interpretation: recommendations from the international consensus on time in range. Diabetes Care. 2019;42(8):1593–603.
25. Amiel SA, Sherwin RS, Simonson DC, Tamborlane WV. Effect of intensive insulin therapy on glycemic thresholds for counterregulatory hormone release. Diabetes. 1988;37(7):901–7.
26. Poff AM, Koutnik AP, Egan B. Nutritional ketosis with ketogenic diets or exogenous ketones: features, convergence, and divergence. Curr Sports Med Rep. 2020;19(7):251–9.
27. Raju B, Cryer PE. Loss of the decrement in intraislet insulin plausibly explains loss of the glucagon response to hypoglycemia in insulin-deficient diabetes: documentation of the intraislet insulin hypothesis in humans. Diabetes. 2005;54(3):757–64.
28. Dagogo-Jack SE, Craft S, Cryer PE. Hypoglycemia-associated autonomic failure in insulin-dependent diabetes mellitus. Recent antecedent hypoglycemia reduces autonomic responses to, symptoms of, and defense against subsequent hypoglycemia. J Clin Invest. 1993;91(3):819–28.
29. Brooks AM, Oram R, Home P, Steen N, Shaw JAM. Demonstration of an intrinsic relationship between endogenous C-peptide concentration and determinants of glycemic control in type 1 diabetes following islet transplantation. Diabetes Care. 2015;38(1):105–12.


Copyright © 2021 by the American College of Sports Medicine