Abstract
Introduction
This year, we screened over 1800 potentially eligible titles on search engines, including PubMed and Google Scholar, between July 1, 2021, and June 30, 2022. We shortlisted 72 original peer-reviewed manuscripts that focused on exercise, nutrition, and diabetes mellitus and ultimately selected 11 papers to represent this research field.
The search this year was packed with new nutrition and exercise-related research with a heavy focus on manipulating the macronutrient composition of the diet and/or adding various micronutrients to the diet for diabetes prevention and management. In the physical activity research space, lifestyle interventions continue to show the beneficial metabolic effects of regular exercise in persons at risk for developing type 2 diabetes and for people already diagnosed with diabetes (type 1 or type 2). Studies continue to refine our knowledge of the appropriate volumes and intensities of exercise for health and metabolic disease treatment. Research also remains focused on how to manage exercise-related dysglycemia using technology-driven solutions.
Key Articles Reviewed
Chiavaroli L, Lee D, Ahmed A, Cheung A, Khan TA, Blanco Mejia S, Mirrahimi A, Jenkins DJA, Livesey G, Wolever TMS, Rahelić D, Kahleová H, Salas-Salvadó J, Kendall CWC, Sievenpiper JL
Zhu R, Larsen TM, Fogelholm M, Poppitt SD, Vestentoft PS, Silvestre MP, Jalo E, Navas-Carretero S, Huttunen-Lenz M, Taylor MA, Stratton G, Swindell N, Drummen M, Adam TC, Ritz C, Sundvall J, Valsta LM, Muirhead R, Brodie S, Handjieva-Darlenska T, Handjiev S, Martinez JA, Macdonald IA, Westerterp-Plantenga MS, Brand-Miller J, Raben A
Hajhashemy Z, Rouhani P, Saneei P
Smith TA, Smart CE, Howley PP, Lopez PE, King BR
Shilo S, Godneva A, Rachmiel M, Korem T, Kolobkov D, Karady T, Bar N, Wolf BC, Glantz-Gashai Y, Cohen M, Zuckerman Levin N, Shehadeh N, Gruber N, Levran N, Koren S, Weinberger A, Pinhas-Hamiel O, Segal E
Cescon M, Choudhary D, Pinsker JE, Dadlani V, Church MM, Kudva YC, Doyle III FJ, Dassau E
Fritsche A, Wagner R, Heni M, Kantartzis K, Machann J, Schick F, Lehmann R, Peter A, Dannecker C, Fritsche L, Valenta V, Schick R, Nawroth PP, Kopf S, Pfeiffer AFH, Kabisch S, Dambeck U, Stumvoll M, Blüher M, Birkenfeld AL, Schwarz P, Hauner H, Clavel J, Seißler J, Lechner A, Müssig K, Weber K, Laxy M, Bornstein S, Schürmann A, Roden M, de Angelis MH, Stefan N, Häring HU
Paldus B, Morrison D, Zaharieva DP, Lee MH, Jones H, Obeyesekere V, Lu J, Vogrin S, La Gerche A, McAuley SA, MacIsaac RJ, Jenkins AJ, Ward GM, Colman P, Smart CEM, Seckold R, King BR, Riddell MC, O'Neal DN
Morrison D, Zaharieva DP, Lee MH, Paldus B, Vogrin S, Grosman B, Roy A, Kurtz N, O'Neal DN
Tagougui S, Legault L, Heyman E, Messier V, Suppere C, Potter KJ, Pigny P, Berthoin S, Taleb N, Rabasa-Lhoret R
Franc S, Benhamou PY, Borot S, Chaillous L, Delemer B, Doron M, Guerci B, Hanaire H, Huneker E, Jeandidier N, Amadou C, Renard E, Reznik Y, Schaepelynck P, Simon C, Thivolet C, Thomas C, Hannaert P, Charpentier G
MACRONUTRIENTS
Effect of Low Glycaemic Index or Load Dietary Patterns on Glycaemic Control and Cardiometabolic Risk Factors in Diabetes: Systematic Review and Meta-Analysis of Randomised Controlled Trials
Chiavaroli L1,2, Lee D1,2, Ahmed A1,2, Cheung A1,2, Khan TA1,2, Blanco Mejia S1,2, Mirrahimi A1,2,3,5, Jenkins DJA1,2,3,4,6, Livesey G7, Wolever TMS1,3,8, Rahelić D9,10,11, Kahleová H12,13, Salas-Salvadó J14,15,16, Kendall CWC1,2,17, Sievenpiper JL1,2,3,4,6
1Department of Nutritional Sciences, Temerty Faculty of Medicine, University of Toronto, Toronto, ON, Canada; 2Toronto 3D Knowledge Synthesis and Clinical Trials Unit, Clinical Nutrition and Risk Factor Modification Centre, St Michael's Hospital, Toronto, ON, Canada; 3Department of Medicine, Temerty Faculty of Medicine, University of Toronto, Toronto, ON, Canada; 4Department of Medical Imaging, Faculty of Medicine, University of Toronto, Toronto, ON, Canada; 5Division of Endocrinology and Metabolism, Department of Medicine, St Michael's Hospital, Toronto, ON, Canada; 6Li Ka Shing Knowledge Institute, St Michael's Hospital, Toronto, ON, Canada; 7Independent Nutrition Logic, Wymondham, UK; 8INQUIS Clinical Research, Toronto, ON, Canada; 9Vuk Vrhovac University Clinic for Diabetes, Endocrinology and Metabolic Diseases, Merkur University Hospital, Zagreb, Croatia; 10School of Medicine, University of Zagreb, Zagreb, Croatia; 11School of Medicine, Josip Juraj Strossmayer University of Osijek, Osijek, Croatia; 12Institute for Clinical and Experimental Medicine, Diabetes Centre, Prague, Czech Republic; 13Physicians Committee for Responsible Medicine, Washington, DC; 14Universitat Rovira i Virgili, Departament de Bioquímica i Biotecnologia, Unitat de Nutrició Humana, Reus, Spain; 15Institut d'Investigació Sanitària Pere Virgili, Hospital Universitari San Joan de Reus, Reus, Spain; 16Consorcio CIBER, MP Fisiopatología de la Obesidad y Nutrición, Instituto de Salud Carlos III, Madrid, Spain; 17College of Pharmacy and Nutrition, University of Saskatchewan, SK, Canada
Background
Previous randomized control trials (RCTs) and meta-analyses have demonstrated that a low glycemic index (GI) or low glycemic load (GL) diet improves glycemia and cardiometabolic risk in people living with diabetes (1,2). This systematic review and meta-analysis, conducted in the spring of 2021, was the first to provide a research grading system on all relevant RCTs for GI/GL diets to help update the European Association for the Study of Diabetes (EASD) clinical practice guidelines for nutrition therapy for people living with diabetes.
Methods
Investigators independently extracted data and assessed risk of bias from published studies (Medline, Embase, Cochrane Library) up until May 13, 2021, and GRADE (Grading of Recommendations Assessment, Development, and Evaluation) was used to assess the certainty of evidence for prescribing three or more weeks of low GI/ GL in type 1 (T1D) or type 2 diabetes (T2D). The primary outcome was glycated hemoglobin (HbA1c) with secondary outcomes for various other markers of glycemic control, lipids, body composition, blood pressure, and inflammation also included.
Results
From a total of 9596 reports published, 29 trials involving a total of 1617 participants with diabetes were included in the final analyses, with most trials conducted in Canada (21%), Australia (17%), and France (10%). Median follow-up duration was 12 weeks; studies often had a crossover design (45% of RCTs), with or without a washout, and a reasonable distribution of women (47%) and men (53%). In general, study participants were middle aged (median age 56 years) and most (90%) were living with T2D, with the remainder having T1D. The median GI values achieved within the intervention and control diets were 49 (range 38–58) and 63 (range 51–86), respectively. (See comment below on what this score is). The median (range) GL achieved in the intervention and control diets were 102 (33–176) and 138 (39–175), respectively. The median percent contribution of energy from carbohydrate in the diets did not differ (49% in the intervention and 48% in the control diets). Most studies (90%) did not initiate a caloric restrictive diet, and most trials were judged as having a low or unclear risk of bias across the GRADE domains. Pooled analyses demonstrated that low GI/GL diets led to a small but clinically meaningful reduction in HbA1c compared with the various control diets (mean difference −0.31% [95% CI, −0.42% to −0.19%]). Low GI/GL diets also showed modest reductions in non-HDL cholesterol (mean difference −0.20 [95% CI, −0.33 to −0.07] mmol/L), LDL cholesterol (−0.17 [95% CI, −0.25 to −0.08] mmol/L), apo B (−0.05 [95% CI, −0.09 to −0.01] g/L), triglycerides (−0.09 [95% CI, −0.17 to −0.01] mmol/L), body weight (−0.66 [95% CI, −0.90 to −0.42] kg), body mass index (−0.38 [95% CI, −0.64 to −0.13]), and small reductions in fasting blood glucose (−0.36 [95% CI, −0.42 to −0.19] mmol/L) and C-reactive protein (−0.41 [95% CI, −0.78 to −0.04] mg/L). The evidence was graded as “high” for the primary outcome (HbA1c improvement), while the evidence for most secondary outcomes was graded as “moderate” owing to downgrades for either inconsistency, imprecision, or evidence of publication bias from small study effects.
Conclusion
This new systematic review and meta-analysis showed that low GI/GL dietary patterns, in comparison with higher GI/GL control diets, provide small but important benefits for glycemic management and other established cardiometabolic risk factors in people living with diabetes over a median follow-up of ∼12 weeks.
Dose-Dependent Associations of Dietary Glycemic Index, Glycemic Load, and Fiber with 3-Year Weight Loss Maintenance and Glycemic Status in a High-Risk Population: a Secondary Analysis of the Diabetes Prevention Study Preview
Zhu R1, Larsen TM1, Fogelholm M2, Poppitt SD3, Vestentoft PS1, Silvestre MP3,4, Jalo E2, Navas-Carretero S5,6,7, Huttunen-Lenz M8, Taylor MA9, Stratton G10, Swindell N10, Drummen M11, Adam TC11, Ritz C1, Sundvall J12, Valsta LM13, Muirhead R14, Brodie S14, Handjieva-Darlenska T15, Handjiev S15, Martinez JA6,7,16,17, Macdonald IA18, Westerterp-Plantenga MS11, Brand-Miller J14, Raben A1,19
1Department of Nutrition, Exercise and Sports, Faculty of Science, University of Copenhagen, Copenhagen, Denmark; 2Department of Food and Nutrition, University of Helsinki, Helsinki, Finland; 3Human Nutrition Unit, School of Biological Sciences, Department of Medicine, University of Auckland, Auckland, New Zealand; 4CINTESIS, Nova Medical School, Universidade Nova de Lisboa, Lisboa, Portugal, 5Centre for Nutrition Research, University of Navarra, Pamplona, Spain, 6Centro de Investigacion Biomedica en Red Area de Fisiologia de la Obesidad y la Nutricion (CIBEROBN), Madrid, Spain; 7IdisNA Instituto for Health Research, Pamplona, Spain, 8Institute for Nursing Science, University of Education Schwäbisch Gmünd, Schwäbisch Gmünd, Germany; 9Division of Physiology, Pharmacology and Neuroscience, School of Life Sciences, Queen's Medical Centre, Nottingham, UK; 10Applied Sports, Technology, Exercise and Medicine (A-STEM) Research Centre, Swansea University, Swansea, UK; 11Department of Nutrition and Movement Sciences, NUTRIM School of Nutrition and Translational Research in Metabolism, Maastricht University, Maastricht, the Netherlands; 12Department of Government Services, Forensic Toxicology Unit, Biochemistry Laboratory, Finnish Institute for Health and Welfare, Helsinki, Finland; 13Department of Public Health Solutions, Finnish Institute for Health and Welfare, Helsinki, Finland; 14School of Life and Environmental Sciences and Charles Perkins Centre, University of Sydney, Sydney, Australia; 15Department of Pharmacology and Toxicology, Medical University of Sofia, Sofia, Bulgaria; 16Department of Nutrition and Physiology, University of Navarra, Pamplona, Spain; 17Precision Nutrition and Cardiometabolic Health Program, IMDEA-Food Institute (Madrid Institute for Advanced Studies), CEI UAM + CSIC, Madrid, Spain; 18Division of Physiology, Pharmacology and Neuroscience, School of Life Sciences, Queen's Medical Centre, MRC/ARUK Centre for Musculoskeletal Ageing Research, ARUK Centre for Sport, Exercise and Osteoarthritis, National Institute for Health Research Nottingham Biomedical Research Centre, Nottingham, UK; 19Steno Diabetes Center Copenhagen, Gentofte, Denmark
Background
Type 2 diabetes (T2D) is thought to be preventable, at least to some degree, in people with prediabetes who undertake lifestyle intervention with regular exercise and caloric restriction. Low-energy diets (LED), with either total or partial meal replacements, are effective in promoting rapid weight loss and improving insulin sensitivity in overweight and obese men and women with prediabetes (3), but weight regain is a common problem after about 1 year of most interventions (4). The PREVIEW (PREVention of diabetes through lifestyle Intervention and population studies in Europe and around the World) project was initiated in 2013 to find the most effective lifestyle (diet and physical activity) approach for the prevention of T2D in overweight and obese participants with increased risk for the disease (5). This 3-year randomized control trial (RCT) with four different intervention arms compared the impact of a high-protein, low-glycemic index (GI) diet with a moderate protein, moderate-GI diet in combination with moderate or high-intensity physical activity levels on the incidence of T2D and its related clinical endpoints (6). The primary goal of this secondary analysis of PREVIEW was to examine the longitudinal and dose-dependent associations of dietary GI, glycemic load (GL), and fiber with body weight and glycemic status during the 3-year weight loss maintenance (WLM) phase of the PREVIEW study participants, irrespective of the original four randomization groups.
Methods
This was a secondary analysis of pooled data from the previously published PREVIEW trial (6) to examine the impact of a low GI diet or a low GL diet or a high fiber diet on weight rebound after study intervention. A total of 1279 participants who were either overweight or obese (age 25–70 years and BMI ≥25 kg·m−2) and living with prediabetes at baseline were included. Multiadjusted linear mixed models with repeated measurements were used to assess longitudinal and dose-dependent associations with various dietary substrategies by merging the participants into one group and dividing them into GI, GL, and fiber tertiles, respectively.
Results
The main findings from this analysis were that each 10-unit increment in GI was associated with a greater regain of weight (0.46 kg/year; 95% CI, 0.23 to 0.68 kg/year; P < .001) and a greater increase in HbA1c level. Each 20-unit increment in GL was also associated with a greater regain of weight (0.49 kg/year; 95% CI, 0.24 to 0.75 kg/year; P<.001) and increase in HbA1c. Compared with those in the lowest tertiles, participants in the highest GI and GL tertiles had significantly greater weight regain and increases in HbA1c. The associations of GI and GL with HbA1c were independent of weight change. Fiber intake was inversely associated with increases in waist circumference but did not associate with weight regain and/or glycemic status.
Conclusion
This secondary analysis of individuals with a high risk of type 2 diabetes from the large international, multiethnic PREVIEW cohort found that higher cumulative average GI and GL intakes are associated with increases in body weight regain after lifestyle intervention and various markers of deteriorated glycemic status. However, fiber intake status, per se, did not appear to impact weight regain or glycemic status.
Comments
According to our 2021–2022 ATTD Yearbook search, several new and important randomized control trials (RCTs) and meta-analyses were conducted to examine the effectiveness of manipulating macronutrient content of a diet to prevent or treat type 2 diabetes (T2D) and to manage glycemic levels in type 1 diabetes (T1D). One common nutritional approach for diabetes management, in general, is to allow relatively liberal carbohydrate intake (i.e., up to 60% of the total energy intake) but change the carbohydrate composition of the diet by decreasing the amount of high glycemic index (GI) foods and increasing the amount of low GI foods consumed in the diet. The GI is a carbohydrate ranking tool that ranks a carbohydrate containing food according to the amount by which it raises blood glucose levels after it is consumed in comparison with reference food such as pure glucose or white bread (7). In this ranking method, a carbohydrate food with a GI of ≤55 is considered low, 56–69 is considered medium, and ≥70 is considered high, based on a glucose scale. The glycemic load (GL) of a food is the GI multiplied by the available carbohydrate (g) in the serving, divided by 100, which is useful since the volume (amount) of carbohydrate also affects the glucose excursion at a meal and the dietary insulin need. For example, an apple that has a GI of 38 and contains ∼15 grams of available carbohydrate has a GL of ∼6 grams.
In the first paper in this section, by Chiavaroli and colleagues, is a new meta-analysis that found that the consumption of low a GI/GL diet, without reducing actual amounts of carbohydrate intake, lowers HbA1c levels by 0.31% in those with diabetes (T1D and T2D studies were pooled). The researchers also demonstrated that there is a significant positive linear dose-response gradient for differences in GL on HbA1c levels in diabetes, showing a net reduction of 0.04% HbA1c units per 10-unit reduction in GL. This analysis is important since some individuals with diabetes cannot adhere to a low carbohydrate diet for very long, even though carbohydrate restriction also lowers HbA1c levels very effectively in both T1D (8) and T2D (9,10), as was shown nicely in several other new papers in this same ATTD yearbook search period. The key is that while both dietary manipulations work, the low carbohydrate diet may be less sustainable for many individuals while the low GI/GL diet may be somewhat more sustainable. Indeed, the second paper, by Zhu et al., showed that once weight loss and improvement in HbA1c are achieved with intensive lifestyle intervention over a period of 8 weeks in persons with prediabetes, as was achieved in the previously published PREVIEW study primary paper (6), it is much easier to sustain that weight loss and improvement in glycemia if foods low in GI and/or low in GL are consumed after the intensive lifestyle intervention ends. This information is important to share, given that the two types of diets consumed in the original PREVIEW study (i.e., high protein and low GI diet vs moderate protein and moderate GI) did not appear to differ in their success for preventing T2D (both dietary types had equal effectiveness). Collectively, all these studies strengthen the notion that carbohydrate quality and quantity are likely critical considerations for lifestyle interventions for prediabetes and T2D and that most of us who are at risk for diabetes, or already have it, should probably be striving toward a diet low in GI foods, particularly if we find that reducing our carbohydrate intake altogether is too difficult to sustain. It should also be noted that using the GI as a rating system for all carbohydrate containing foods comes with risk, since it tends to make people believe that foods are either “good” or “bad” for their diabetes (or health), which is an oversimplification (11).
MICRONUTRIENTS AND NUTRACEUTICALS
Dietary Calcium Intake in Relation to Type-2 Diabetes and Hyperglycemia in Adults: a Systematic Review and Dose-Response Meta-Analysis of Epidemiologic Studies
Hajhashemy Z1,2, Rouhani P1,2, Saneei P2
1Students' Research Committee, Isfahan University of Medical Sciences, Isfahan, Iran; 2Department of Community Nutrition, School of Nutrition and Food Science, Isfahan University of Medical Sciences, Isfahan, Iran
Background
While several modifiable and nonmodifiable risk factors are thought to be linked to the development of type 2 diabetes (T2D), including age, sex, genetic background, race/ethnicity, body weight, physical activity levels, and dietary patterns, some debate exists for the role of micronutrients in the disease etiology. One nutrient that gets regular attention for its possible link with T2D development is dietary calcium (Ca) (12 –14). The aim of this study was to conduct a systematic review and meta-analysis to determine whether higher dietary calcium intake is associated with a lower risk of T2D or hyperglycemia in adults.
Methods
The investigators examined all published articles found in MEDLINE (Pubmed), Web of Science, and Scopus electronic databases as well as in Google Scholar up to May 2021. All published papers were included in the meta-analysis if they (1) had a cohort, cross-sectional, or case-control design; (2) investigated an adult population (≥18 years), regardless of their health status; (3) considered dietary Ca intake as the exposure and reported the risk for abnormal glucose homeostasis, including T2D, prediabetes or hyperglycemia as the outcomes of interest; (4) reported relative risks (RRs), hazard ratios (HR), or odds ratios (ORs), with 95% confidence intervals (CIs) for the association of dietary calcium intake and abnormal glucose homeostasis.
Results
Approximately 4000 papers were initially screened, with ∼100 reports retrieved in full for further screening; 17 eligible studies were included in the final meta-analysis (8 prospective, 9 cross-sectional). The assessment of dietary Ca intake was performed using food frequency questionnaires (FFQs) in 12 reports, food recall in four investigations, and food record in one study. Nine of the included studies considered T2D as the outcome, and the others reported high blood glucose (HBG) or hyperglycemia (fasting blood glucose [FBG] ≥ 100 or 110 mg/dL) as the outcome of interest. Among the 17 studies, 12 were deemed as “high-quality” (score of 8 or more out of a 10-point scale), while the remaining studies were classified as “low-quality”. Results from 7 cohort studies (representing 255,744 adults) showed that individuals in the highest category of Ca intake, in comparison with the lowest category of Ca intake, had 18% lower risk of developing T2D (RR, 0.82; 95% CI, 0.74 to 0.92) but the heterogeneity between studies was deemed as “moderate” (I2 = 53.6, PQ-test = 0.02). Interestingly, the Ca intake—T2D relation was significant only in Asian countries (RR, 0.75; 95% CI, 0.69 to 0.82) and there was no significant relation in non-Asian regions (RR, 0.98; 95% CI, 0.87 to 1.11). Among participants with a mean age ≥50 years and those in developing countries, higher dietary Ca intake was protectively associated with lower risk of T2D in subgroups of women, high-quality studies, and investigations with an adjustment for magnesium intake. Based on eight studies that had a dose response examination, a 300, 600, and 1000 mg/day increase in dietary Ca intake was significantly related to a 7%, 13%, and 20% decrease in risk of T2D (RR, 0.93; 95% CI 0.89 to 0.98), (RR, 0.87; 95% CI 0.79 to 0.97) and (RR, 0.80; 95% CI 0.67 to 0.95), respectively. There was no significant relation between dietary Ca intake above 750 mg/d and risk of developing T2D, however.
Conclusion
This new meta-analysis demonstrates that increasing Ca intake from less than 300 mg/day to amounts up to ∼750 mg/day is associated with a stepwise reduction in the relative risk for developing T2D in adults over the age of 50 years.
Comments
Several papers this year focused on fortifying diets with micronutrients like vitamins, minerals, and nutraceuticals (e.g., dairy products, certain types of long chain fatty acids, food extracts, etc.). A nutraceutical is defined as any substance that is a whole food, or part of a food, that has a beneficial physiologic effect and provides medicinal or health benefits, including the prevention and treatment of various diseases like cancer, heart disease, Alzheimer disease, or diabetes. In papers published this year, for example, all the following dietary supplements were shown to influence either diabetes development or glucose management in persons with diabetes: linoleic acid (15); coffee (16); whey protein (17); ginger (18); saffron (19,20); omega 3 fatty acids (21,22); Morus Alba leaf extract (23); vitamin K1 (24); branched chain amino acids (25); red raspberry (26); and Artemisia vulgaris (a type of flowering plant also known as mugwort) (27). We selected the paper reviewed in this section from this long list of articles on potential dietary supplements because Ca intake levels in adults have long been linked to changing risk for a number of chronic disease processes, including type 2 diabetes, but excessive Ca intake (without coadministered vitamin D) might also increase the risk for myocardial infarction (28). A previous meta-analysis demonstrated a possible link between Ca consumption and T2D development, but this relationship was confounded by magnesium intake (29). This latest meta-analysis shows clearly that an inverse association exists between dietary Ca intake and risk of T2D in the general adult populations in a dose-response manner, with the intake of 750 mg/day sufficient to reduce the risk of developing T2D. It is also an important finding that more is not better for diabetes prevention, since the consumption of >900 mg/d appears to increase the risk for cardiovascular disease mortality and all-cause mortality (30). In addition, Ca is also important for bone health in diabetes (31). So, for now, we should all try to achieve this ∼750 mg/day threshold for calcium intake in some way, whether through supplements, plant-based foods, animal-based foods, or some combination of these. A list of foods enriched in calcium can be found here:
INSULIN DOSING STRATEGIES FOR COMPLEX MEALS
For a High Fat, High Protein Breakfast, Preprandial Administration of 125% of the Insulin Dose Improves Postprandial Glycaemic Excursions in People with Type 1 Diabetes Using Multiple Daily Injections: a Cross-Over Tria
Smith TA1,2, Smart CE1,2,3, Howley PP4, Lopez PE1,2,3, King BR1,2,3
1Faculty of Health and Medicine, University of Newcastle, Callaghan, Australia; 2Hunter Medical Research Institute, New Lambton Heights, Australia; 3Department of Paediatric Endocrinology, John Hunter Children's Hospital, New Lambton Heights, Australia; 4Faculty of Science, University of Newcastle, Callaghan, Australia
Background
A major challenge for individuals living with type 1 diabetes (T1D) is controlling glycemia after meals that have significant amounts of carbohydrates and are mixed with high fat or high protein foods (32). Adding protein and fat to meals with carbohydrate increases peak glucose levels and extends the glucose profile (32,33), suggesting that more insulin may be needed at meal time and perhaps later as well (32). To increase the time action of prandial insulin for individuals on insulin pumps, a modified bolus profile can be given (e.g., dual wave or square wave bolus), but those on multiple daily injections (MDI) of insulin do not have this ability. For the latter group, the use of regular insulin may be advantageous in these situations, since the time action profile is much longer for regular insulin than for rapid-acting insulin analogues.
Methods
Individuals with T1D (N=24, ages 9–35 years old) using MDI therapy consumed a high fat, high protein “test” breakfast. Researchers compared preprandial insulin-dosing based on the study participant's unique insulin-to-carbohydrate ratio (ICR) with the following modifications: 1) 100% of the usual ICR using aspart insulin before the meal (100Asp); 2) 125% of the usual ICR using aspart insulin before the meal (125Asp); 3) 125% of the usual ICR using regular insulin before the meal (125Reg); and 4) 125% of the usual ICR using aspart insulin, but splitting the dose of aspart by administering 100% before and 25% one hour after the meal (100:25Asp). Postprandial sensor glucose (Dexcom G5) was measured for 5 hours after a meal.
Results
The 5-hour incremental areas under the curve (AUCs) for 100Asp, 125Asp, 125Reg and 100:25Asp were 620 mmol/L·min (95% CI, 451 to 788), 341 mmol/L·min (95% CI, 169 to 512), 675 mmol/L·min (95% CI, 504 to 847), and 434 mmol/L·min (95% CI, 259 to 608), respectively. The 5-hour incremental AUC for 125Asp was significantly lower than for 100Asp (P=.016) and for 125Reg (P=.002). There was only one episode of hypoglycemia in the 96 meal sessions (in 125Reg).
Conclusion
For a high fat (40 g), high protein (50 g) meal that also contains carbohydrate (30 g), giving 125% of the usual ICR as aspart insulin 30 minutes before the meal improves postprandial glycemia significantly more than giving the usual aspart insulin dose (100Asp). Giving insulin in a split dose of aspart (100:25Asp) or replacing aspart with regular insulin provides no glycemic benefit over this newly recommended approach or the usual apart dose approach (100Asp).
Comments
Many of us living with T1D have complex mixed meals that contain all the macronutrients in a similar amount. These energy dense meals are sometimes called “high fat” or “high protein” meals. For people on standard insulin pump therapy, these meals can be managed by either dual-wave or square wave boluses, but this option is unavailable for those on MDI therapy. For these individuals, this small proof of concept study shows that meals with significant amounts of protein (∼50 g) and fat (∼40 g) can be managed by giving 125% of the usual ICR using aspart insulin. This approach is simpler and possibly even more effective than giving a split dose of aspart or trying to use regular insulin for these types of meals. Sometimes, simpler can be better!
MEAL AND EXERCISE PREDICTION STUDIES
Prediction of Personal Glycemic Responses to Food for Individuals with Type 1 Diabetes Through Integration of Clinical and Microbial Data
Shilo S1,2,3, Godneva A1,2, Rachmiel M4,5, Korem T1,2,6, Kolobkov D1,2, Karady T1,2, Bar N1,2, Wolf BC1,2, Glantz-Gashai Y3, Cohen M3,7, Zuckerman Levin N,3,7 Shehadeh N3,7, Gruber N5,8, Levran N8,9, Koren S5,10, Weinberger A1,2, Pinhas-Hamiel O5,8, Segal E1,2
1Department of Computer Science and Applied Mathematics, Weizmann Institute of Science, Rehovot, Israel; 2Department of Molecular Cell Biology, Weizmann Institute of Science, Rehovot, Israel; 3Pediatric Diabetes Clinic, Institute of Diabetes, Endocrinology and Metabolism, Rambam Health Care Campus, Haifa, Israel; 4Pediatric Endocrinology Unit, Shamir Medical Center, Zerifin, Israel; 5Sackler School of Medicine, Tel Aviv University, Tel Aviv, Israel; 6Department of Systems Biology, Columbia University, New York, NY; 7Bruce Rappaport Faculty of Medicine, Technion, Haifa, Israel; 8Pediatric Endocrine and Diabetes Unit, Edmond and Lily Safra Children's Hospital, Sheba Medical Center, Ramat-Gan, Israel; 9Robert H. Smith Faculty of Agriculture, Food and Environment, The Hebrew University of Jerusalem, Rehovot, Israel; 10Diabetes Unit, Shamir Medical Center, Zerifin, Israel
This manuscript is also discussed in DIA-2023-2508, page S-118.
Background
Several factors are thought to influence the bolus insulin needs for meals for individuals with type 1 diabetes (T1D), and postprandial glycemic responses (PPGRs) to certain meals remains a major challenge. This study was designed to determine the main contributors to meal excursions in adults and children with T1D to help develop new machine learning algorithms for PPGR predictions.
Methods
One-hundred and twenty-one individuals (75 adults, 46 children) with T1D, who all wore continuous glucose monitors (CGM) and continuous subcutaneous insulin infusion (CSII) devices, captured real-time dietary intake using a designated mobile app for 2 weeks. PPGR was quantified, and a machine learning algorithm was developed for PPGR prediction using CGM data, insulin dosing, dietary habits, blood parameters, participant characteristics, exercise, and gut microbiota data. Data of the PPGR of 900 healthy control participants and 41,371 meals were also integrated into the model. The performance of the models was evaluated with cross validation.
Results
The PPGR to 6377 isolated meals (i.e., >90 min from an adjacent meal, some meals were merged if there was <30 minutes in between two meals and the first meal was >50 calories) from the 121 T1D participants were collected. Average carbohydrate, fat, and protein consumption was 39.1±8%, 41.6±10.2%, and 17.1±4.3% of total energy, respectively. Test meals (i.e., standardized breakfasts that ranged from 60 g glucose to 65 g bread +20 g butter) that were provided by the study team had high reproducibility within participants when the baseline glucose was also similar (r=0.63). In real-life meals, however, the prediction of PPGR using a model that is based solely on the meal's carbohydrate content achieved a relatively low correlation with PPGRs (r=0.16) and explained only ∼3% of the variance in the glycemic response. Including baseline glucose and insulin bolus amount improved the prediction but still explained only ∼16% of the variance in the glycemic response. Finally, a more complete model, with integration of glucose measurements, insulin dosages, dietary habits, blood parameters, anthropometrics, exercise, and gut microbiota achieved a higher correlation for the PPGRs of meals and increased the explained variance to ∼35% (r=0.59). Adding the PPGR data from the meals consumed by the individuals without diabetes did not improve model prediction; however, several other variables could improve the predictive capacity of the model, including premeal glucose, glucose trend in the 30 minutes prior to the meal, meal carbohydrate content, the ratio between carbohydrate and fat in the meal, time of day, and microbial composition of a stool sample.
Conclusion
A newly generated prediction model for the glycemic response to real-life meals that uses a comprehensive clinical and microbiome profile as input helps predict the PPGR to meals in adults and children with T1D.
Activity Detection and Classification from Wristband Accelerometer Data Collected on People with Type 1 Diabetes in Free-Living Conditions
Cescon M1, Choudhary D2, Pinsker JE 3 , Dadlani V 4 , Church MM 5 , Kudva YC 4 , Doyle III FJ1, Dassau E6
1Harvard John A. Paulson School of Engineering and Applied Sciences, Harvard University, Boston, MA (now at University of Houston, TX); 2Harvard John A. Paulson School of Engineering and Applied Sciences, Harvard University, Boston, MA (now at University Oxford); 3Sansum Diabetes Research Institute, Santa Barbara, CA (now at Tandem Diabetes Care, Inc); 4Mayo Clinic, Rochester, MN; 5Sansum Diabetes Research Institute, Santa Barbara, CA; 6Harvard John A. Paulson School of Engineering and Applied Sciences, Harvard University, Boston, MA (now with Eli Lilly and Company)
Background
Different types, intensities, and durations of exercise can all cause disturbances to blood glucose concentrations for individuals with type 1 diabetes (T1D). In this study, a wrist-worn accelerometer was used to detect and estimate physical activity and sedentary behaviors in adults with T1D. The goal of this study is to use this information to better inform automated insulin delivery (AID) system needs around exercise.
Methods
A total of 20 adults with T1D were asked to wear the Empatica E4 wristband for 5 weeks on their nondominant wrist. In a personal logbook, all participants reported the time, duration, type, and intensity of any sedentary, household, lifestyle, sport, or gym activities. The intensity was graded on a scale of 0 to 3 in which 0 was sedentary, 1 was mild, 2 was moderate, and 3 was vigorous activity. This study was conducted in free-living conditions with no predefined protocol to allow for accurate representation of daily living. Two methods were used for activity intensity classification: Method 1, Random Forest (RF) classifier machine learning method, and Method 2, a threshold-based algorithm.
Results
The results are reported for all participants that performed at least two different activities on at least 2 separate days and annotated the activities in their logbooks. Clinical investigators grouped participants based on fitness level. For Method 1, the RF classifier was tested on target classes of sedentary, mild, moderate, and vigorous activities. The average accuracy for activity level categorization was 99.99%. For Method 2, sedentary behavior was annotated in 32 instances and predicted with 96.87% precision and 96.87% recall. Mild activity level was annotated in 48 instances and predicted with 74.50% precision and 79.17% recall. Moderate activity level was annotated in 95 instances and predicted with 88.29% precision and 87.36% recall. And finally, vigorous activity level was annotated in 15 instances and predicted with 100% precision and 86.67% recall. The evaluation of accuracy in predicting the intensity level for the most common activities reported accuracy levels of 68.5% for walking, 100% for hiking, 90.2% for running, 83.3% for yard work, 100% for housework, and 50.0% for high intensity interval training.
Conclusion
This work demonstrates that in individuals with T1D, physical activity and sedentary behaviors can be detected in free-living conditions and further classified as “mild”, “moderate” or “intense” with good accuracy and precision using a triaxial wrist-worn accelerometer.
Comments
The first paper, by Shilo et al., aimed to determine the main contributors to meal excursions in individuals with T1D to help develop new machine learning algorithms for predicting postprandial glucose excursions in T1D. They found, somewhat surprisingly, that the meal's carbohydrate content was not that important in the meal-related glucose excursion in their prediction model. This is odd since we intuitively know that meals with high amounts of carbohydrates, particularly if they have a high glycemic index, tend to cause glucose to rise well above target even if insulin is administered accordingly. The reason why the carbohydrate content of a meal did not correlate well with the PPGR in their model is likely threefold. First, an appropriate insulin bolus was given for each of the standardized test meals based on the carbohydrate content of that meal using the test subject's unique insulin-to-carbohydrate ratio (ICR). As such, the amount of carbohydrate of the meal was already “covered” by the insulin. Obviously, if bolus insulin was not given at the standardized meal, the amount of PPGR would be highly correlated to the volume of carbohydrate consumed. Second, the amount of carbohydrate given was not that large (only up to 30 grams). Third, it is well known that several factors influence the glycemic excursion of a meal, including the macronutrient mix, the glycemic index of the carbohydrates consumed, baseline glucose level, and the individual's ever changing insulin sensitivity based on activity levels, time of day, stress levels, etc. What we learned from this study is that several factors appear to be associated with greater meal excursions, including higher carbohydrate meals, adding a lot of fat to high carbohydrate meals, morning meals, glucose levels elevated at baseline or trending upward at baseline, and being inactive before after meals. Future studies using this model prediction approach to help improve insulin dosing for certain meals in T1D are warranted.
The second paper, by Cescon et al., presents two methods to detect and grade physical activity using a three-axis accelerometer wrist-worn device. What sets this manuscript apart from many of the other studies that have focused on meal and/or exercise detection is that Cescon and colleagues conducted this data collection in unsupervised free-living conditions without specific protocol designs and guidance on what activities participants should perform. Two key measures from this study that may further enhance current automated insulin detection systems for exercise include the activity intensity signal (i.e., whether the activity is aerobic or anaerobic) and the time spent in each activity level. We feel that these are critical elements to know when determining insulin dosing strategies for exercise (34).
LIFESTYLE INTERVENTION STUDIES TO PREVENT TYPE 2 DIABETES
Different Effects of Lifestyle Intervention in High- and Low-Risk Prediabetes: Results of the Randomized Controlled Prediabetes Lifestyle Intervention Study (PLIS)
Fritsche A1,2,3, Wagner R1,2,3, Heni M1,2,3, Kantartzis K1,2,3, Machann J1,3,5, Schick F1,4, Lehmann R1,3,5, Peter A1,3,5, Dannecker C1,3, Fritsche L1,3, Valenta V1,3, Schick R1, Nawroth PP1,6,7,8, Kopf S1,6, Pfeiffer AFH4,9, Kabisch S1,9, Dambeck U1,9, Stumvoll M1,10, Blüher M1,10, Birkenfeld AL1,11, Schwarz P1,11, Hauner H1,12, Clavel J1,12, Seißler J1,13, Lechner A1,13, Müssig K1,14,15, Weber K1,15, Laxy M1,16, Bornstein S1,11, Schürmann A1,9, Roden M1,14,15, de Angelis MH1,17,18, Stefan N, Häring HU1,2,3.
1German Center for Diabetes Research (DZD), Neuherberg, Germany; 2Division of Diabetology, Endocrinology and Nephrology, Department of Internal Medicine IV, Eberhard-Karls University Tübingen, Tübingen, Germany; 3 Institute for Diabetes Research and Metabolic Diseases, Helmholtz Center Munich, University of Tübingen, Tübingen, Germany; 4Section on Experimental Radiology, Department of Radiology, University of Tubingen, Tubingen, Germany; 5Institute for Clinical Chemistry and Pathobiochemistry, Department for Diagnostic Laboratory Medicine, University Hospital of Tübingen, Tübingen, Germany; 6Department of Medicine I and Clinical Chemistry, University Hospital of Heidelberg, Heidelberg, Germany; 7Institute for Diabetes and Cancer, IDC Helmholtz Center, Munich, Germany; 8Joint Heidelberg-IDC Translational Diabetes Program, Neuherberg, Germany; 9German Institute of Human Nutrition Potsdam-Rehbrücke, Nuthetal, Germany; 10Department of Medicine, Endocrinology and Nephrology, Universität Leipzig, Leipzig, Germany; 11Department of Internal Medicine III, Technische Universität Dresden, Dresden, Germany; 12Institute of Nutritional Medicine, School of Medicine, Technical University of Munich, Munich, Germany; 13Diabetes Research Group, Medical Department 4, Ludwig-Maximilians University Munich, Munich, Germany; 14Division of Endocrinology and Diabetology, Medical Faculty, Heinrich Heine University Düsseldorf, Düsseldorf, Germany; 15Institute for Clinical Diabetology, German Diabetes Center, Leibniz Center for Diabetes Research at Heinrich Heine University Düsseldorf, Düsseldorf, Germany; 16Institute of Health Economics and Health Care Management, Neuherberg, Germany; 17Institute of Experimental Genetics, IEG Helmholtz Center Munich, Neuherberg, Germany; 18Experimental Genetics, School of Life Sciences, Weihenstephan, Technical University of Munich, Munich, Germany
Background
Although lifestyle interventions are effective in reducing the risk for type 2 diabetes (T2D) in adults at risk, it appears that they do not work for everyone (35). In fact, every fifth patient of the lifestyle intervention group in the Diabetes Prevention Program (DPP) developed T2D within 4 years (36), and there are some people with prediabetes who did not progress to diabetes in the 11-year follow-up even without intervention (37). This suggests that some people with prediabetes may be less responsive to lifestyle intervention and while others may be more responsive. This randomized control trial (RCT) determined if high-risk individuals, who are deemed “nonresponders” to lifestyle intervention based on measures of body weight changes and certain glycemic markers, respond more favorably if the lifestyle intervention is intensified.
Methods
Included in this study were 1105 adults with prediabetes, who were then stratified into a high-risk (HR) or low-risk (LR) group for T2D based on an oral glucose tolerance test (OGTT) and liver fat content. HR participants were characterized by a reduced insulin secretion (based on the disposition index [DI]) and/or insulin resistance (based on the sensitivity index [ISI]) with elevated liver fat content. LR individuals were randomly assigned to conventional lifestyle intervention (LI) according to the DPP protocol or control (1:1) while HR individuals were also assigned to conventional or intensified LI with doubling of the required exercise volume (1:1) for a 12-month period (i.e., four groups, a total of 1105 individuals in study). Participants of the intensified LI group received 16 coaching sessions in total over 1 year with advice to exercise 6 hours weekly, rather than the 3 hours weekly that was advised to the usual care groups. The primary outcome measure, 2-hour post-load glucose concentration, was measured at 6 months and 12 months. Secondary outcomes included markers of insulin secretion and sensitivity during the OGTT and liver fat content before and after interventions.
Results
Post-challenge glucose in the OGTT decreased in all four groups, but the more intensively treated HR group had a greater improvement in OGTT than the conventionally treated HR group (mean difference −0.29 mmol/L [95% CI, −0.54 to −0.04]). The secondary outcomes of liver fat content and cardiovascular risk profile were also more favorable in the more intensively treated HR group than in the conventionally treated HR group. During a follow-up of 3 years, intensified LI had a higher probability of normalizing glucose tolerance than did conventional LI.
Conclusion
This RCT demonstrates that HR individuals with prediabetes can improve their glycemic and cardiometabolic outcomes by doubling the weekly exercise volume of the current recommended lifestyle intervention. Thus, individualized risk assessment, using an OGTT and liver fat assessment, may be useful to identify HR phenotypes that can benefit from more intensive exercise counseling.
Comments
Researchers who conduct exercise intervention studies often use the term “nonresponders” when they refer to individuals that do not derive the expected benefits of regular exercise (38). In some individuals with high risk for developing T2D, adding the typical “prescription” of exercise at a “dose” of ∼180 minutes per week does not prevent T2D (36,39), so these individuals may be nonresponders. This new study shows that high-risk individuals, as identified by OGTT and liver fat content, do indeed respond favorably from a cardiometabolic perspective if the exercise prescription is increased to ∼360 minutes per week (i.e., 6 hours of training per week). Admittedly, this is a lot of exercise for many individuals with prediabetes who may be unaccustomed to regular exercise. This amount of exercise was facilitated by exercise coaches and only to those who could do at least 3 hours per week at the start of the lifestyle program. This was a relatively short study, however (only 12 months), and there was some attrition in the HR intensified lifestyle treatment arm (18%). Further studies with longer intervention times are needed to determine if more aggressive exercise prescriptions are needed for some individuals at risk for T2D who do not appear to respond favorably to lifestyle intervention.
AUTOMATED INSULIN DELIVERY, TECHNOLOGY, AND ADDITIONAL METRICS FOR EXERCISE
A Randomized Crossover Trial Comparing Glucose Control During Moderate-Intensity, High-Intensity, and Resistance Exercise with Hybrid Closed-Loop Insulin Delivery While Profiling Potential Additional Signals in Adults with Type 1 Diabetes
Paldus B1,2, Morrison D1, Zaharieva DP3, Lee MH1,2, Jones H1,2, Obeyesekere V2, Lu J1,2, Vogrin S1, La Gerche A4.5, McAuley SA1,2, MacIsaac RJ1,2, Jenkins AJ1,6, Ward GM1, Colman P7, Smart CEM8, Seckold R8, King BR8, Riddell MC3, O'Neal DN1,2
1Department of Medicine, University of Melbourne, Melbourne, Victoria, Australia; 2Department of Endocrinology and Diabetes, St Vincent's Hospital Melbourne, Melbourne, Victoria, Australia; 3School of Kinesiology and Health Science, Muscle Health Research Centre, York University, Toronto, Ontario, Canada; 4Department of Cardiology, St Vincent's Hospital Melbourne, Melbourne, Victoria, Australia; 5Clinical Research Domain, Baker Heart and Diabetes Institute, Melbourne, Victoria, Australia; 6NHMRC Clinical Trials Centre, University of Sydney, Camperdown, New South Wales, Australia; 7Department of Diabetes and Endocrinology, Royal Melbourne Hospital, Parkville, Victoria, Australia; 8John Hunter Children's Hospital, Newcastle, New South Wales, Australia
Background
For many individuals with type 1 diabetes (T1D), exercise and physical activity continue to challenge automated insulin delivery (AID) systems, and the risk of hypoglycemia is still present. The aim of this study was to compare glycemia using AID during three exercise modalities: moderate-intensity exercise (MIE), high-intensity exercise (HIE), and resistance exercise (RE).
Methods
This multisite randomized crossover trial was conducted on 30 adults with T1D (16 men, 14 women). Participants completed at least a 1-week run-in period with the Medtronic MiniMed 670G system and glucose sensor with a target glucose of 120 mg/dL. Participants were randomized to 40 minutes of MIE, HIE, or RE with a temporary target set 2 hours prior to each exercise condition. If glucose levels were <126 mg/dL, 15 grams of carbohydrates were consumed before exercise to reduce the risk of hypoglycemia during the activity. The primary outcome was continuous glucose monitoring (CGM) percent time-in-range (TIR) for 14 hours after exercise. The secondary outcomes included changes in plasma glucose, ketones, lactate, plasma insulin, counterregulatory hormones, and accelerometer data.
Results
The median TIR 0 to 14 hours after exercise was 81% (IQR, 67%, 93%) for HIE, 91% (IQR, 80%, 94%) for MIE, and 80% (IQR, 73%, 89%) for RE, and there were no differences between exercise types (MIE vs HIE, P=.11; MIE vs. RE, P=.11; and HIE vs. RE, P=.90). Time below range (TBR; <70 mg/dL) was <1% for all exercise conditions. Noradrenaline (P=.01 and P=.004), cortisol (P<.001 and P=.001), lactate (P≤.001 and P≤.001), and max heart rate (P=.007 and P=.001) were significantly greater during HIE and RE than during MIE. Growth hormone increased significantly more during HIE than during MIE (P=.024).
Conclusion
In a highly controlled exercise setting in which the timing of exercise was fixed, there were no clinically significant differences in glycemia among HIE, MIE, and RE when using the MiniMed 670G system in adults with T1D. Counterregulatory hormones, lactate, and accelerometry were able to differentiate the type and intensity of exercise and may be important additional signals to help inform insulin needs during exercise.
Comparable Glucose Control with Fast-Acting Insulin Aspart Versus Insulin Aspart Using a Second-Generation Hybrid Closed-Loop System During Exercise
Morrison D1, Zaharieva DP2, Lee MH1,3, Paldus B1,3, Vogrin S1, Grosman B4, Roy A4, Kurtz N4, O'Neal DN1,3
1Department of Medicine, University of Melbourne, Melbourne, Australia; 2Division of Endocrinology, Department of Pediatrics, School of Medicine, Stanford University, Stanford, CA; 3Department of Endocrinology and Diabetes, St Vincent's Hospital Melbourne, Melbourne, Australia; 4Medtronic Diabetes, Northridge, CA
This manuscript is also discussed in DIA-2023-2504, page S-44.
Background
Closed-loop systems have been shown to improve glycemia in individuals with type 1 diabetes (T1D); however, systems continue to be challenged in their ability to address rapid changes in glycemia associated with exercise. The present study aimed to compare the effects of faster-acting insulin aspart (FiAsp) and insulin aspart on glucose management and performance of the new MiniMed Advanced Hybrid Closed Loop (AHCL) system during and after exercise (moderate-intensity and high-intensity).
Methods
This was a randomized crossover study comparing FiAsp with insulin aspart across four moderate-intensity exercise (MIE) and high-intensity exercise (HIE) bouts using AHCL. Participants were randomly assigned to FiAsp and insulin aspart each for 6 weeks. During each insulin therapy period, 40 minutes of MIE (∼50% of VO2max) and HIE (∼80% of VO2max) were performed, in random order, at least 4 hours after any meal- or correction-related bolus (i.e., low or no insulin on board). A higher glucose target (temporary target, 8.3 mmol/L) was set 2 hours before exercise for all exercise conditions, and supplemental carbohydrate (15 g), without an insulin bolus, was administered 15 minutes before exercise if the glucose level was <7.0 mmol/L. For all exercise interventions, the temporary glucose target was ceased immediately after exercise with reversion to the usual target of 5.6 mmol/L. The primary outcome was continuous glucose monitoring (CGM) time-in-range (TIR, 3.9–10.0 mmol/L) for 24 hours after exercise.
Results
A total of 16 adults (9 men; age 48 [IQR, 37, 57] years; hemoglobin A1c 7.0 [IQR, 6.4, 7.2]%; duration diabetes 30 [IQR, 17, 41] years) were recruited for the study. During both insulin formulation arms of the study (FiAsp vs insulin aspart), there were no significant differences between exercise intensity (MIE vs HIE). The CGM median TIR 24 hours after exercise was >81%, percent time below range (<3.9 mmol/L) was <4%, and percent time above range (>10.0 mmol/L) was <17% for both exercise conditions (MIE vs HIE) and insulin formations, with no significant differences between insulins (P>.05). In the 2 hours after exercise and overnight, the TIR approached 100% for all conditions.
Conclusion
In a controlled setting, when comparing FiAsp versus insulin aspart in a second generation closed-loop (i.e., MiniMed AHCL) system during and after exercise, both insulin formulations performed very well (80% to 100% TIR). Overall, this study found that for adults with T1D performing exercise using AHCL technology, FiAsp did not demonstrate a clinical advantage over insulin aspart.
Comments
Even with advancements in automated insulin delivery (AID) systems for youth and adults with T1D, one area that continues to cause disturbances to glycemia is exercise. Consensus guidelines and position statements recommend significant preplanning for exercise to maintain glucose levels during and after exercise (40). These guidelines include setting higher glucose targets 1 to 2 hours before exercise, reducing insulin on board, and, if necessary, consuming carbohydrates to reduce the risk of hypoglycemia. A primary reason for these recommendations of preplanning for exercise is due to the slow pharmacokinetics and pharmacodynamics of subcutaneous insulin delivery. Unfortunately, for many individuals with T1D, significant preplanning for exercise is not always a practical or sustainable approach. As such, there has been a heightened focus on automating meal and exercise detection for AID systems. In addition, integrating additional signals such as ketone, lactate, and/or heart rate metrics into AID systems may be a strategy to better anticipate the effects of exercise.
The first paper in this section, by Paldus et al., found that AID using the MiniMed 670G system led to satisfactory glucose management during the 14-hour period following moderate- or high intensity- or resistance-type exercise. Of note, in this controlled study setting, participants set the higher glucose target (8.3 mmol/L) 2 hours before all exercise interventions. Additional hormonal and metabolic signals were also measured as potential candidates for signaling the onset of various modalities of exercise. Although noradrenaline, cortisol, lactate, and heart rate increased with exercise, the potential utility of these signals to modulate insulin dosing in AID systems continues to be limited by the pharmacokinetics of subcutaneous insulin delivery. As such, the second paper, by Morrison et al., aims to get to the root of the problem by assessing glucose management and the performance of a second-generation AID system during exercise with FiAsp insulin compared with standard insulin aspart. This study hoped to highlight the faster onset and offset of insulin action with FiAsp during exercise. Unfortunately, there was no clinically meaningful benefit for glucose management during and after exercise when comparing the two insulin formulations in adults with T1D. This study has similar outcomes to the results found by Dovc et al. (41) that demonstrated FiAsp in a first-generation AID system during exercise in young adults with T1D to be safe and as effective but not superior to insulin aspart. So, does this mean it is time to give up on new insulin formulations? Well, no. In fact, FiAsp has been shown to be superior to insulin aspart when used with a second-generation AHCL system for postprandial glucose management (42). And with newer fast-acting insulin formulations (e.g., Lyumjev, etc.) now available in some countries, we can expect the next round of research studies to be testing the safety and efficacy of these formulations, particularly around meals and exercise with AID systems.
HYPOGLYCEMIA PREVENTION STRATEGIES AND INTERACTIONS WITH INSULIN
Anticipated Basal Insulin Reduction To Prevent Exercise-Induced Hypoglycemia in Adults and Adolescents Living with Type 1 Diabetes
Tagougui S1,2,3, Legault L1,4, Heyman E3, Messier V1, Suppere C1, Potter KJ1, Pigny P5, Berthoin S3, Taleb N1,6,7, Rabasa-Lhoret R1,2,7
1Montreal Clinical Research Institute (IRCM), Montreal, Canada; 2Département de nutrition, Université de Montréal, Montreal, Canada; 3Univ. Lille, Univ. Artois, Univ. Littoral Côte d'Opale, ULR 7369 - URePSSS - Unité de Recherche Pluridisciplinaire Sport Santé Société, Lille, France; 4McGill University Health Center (MUHC), Montreal Children's Hospital, Montreal, Canada; 5Laboratoire de Biochimie-Hormonologie, CHU Lille, Centre de Biologie-Pathologie, Lille, France; 6Division of Biomedical Sciences, Faculty of Medicine, Université de Montréal, Montréal, Canada; 7Endocrinology Division, Montreal Diabetes Research Center (MDRC), Montreal University Hospital (CHUM), Montreal, Canada
Background
For many individuals with type 1 diabetes (T1D), exercise-associated hypoglycemia remains a challenge for glucose management. This is primarily due to the body's inability to reduce exogenous circulating insulin in a timely fashion. This study assessed the effect of two different timings for basal insulin rate reduction (one early, one later) on a standard (i.e., open loop) insulin pump on glucose changes and the association between circulating insulin levels and muscle vasoreactivity.
Methods
A total of 20 individuals (10 adults, 10 adolescents) with T1D using continuous subcutaneous insulin infusion (CSII) completed the study. In random order, all participants completed 60 minutes of moderate-intensity cycling with an 80% basal insulin rate reduction set 40 minutes (T-40) or 90 minutes (T-90) before the onset of exercise. Muscle hemodynamics were monitored during exercise using near-infrared spectroscopy.
Results
An 80% basal insulin reduction 90 minutes before exercise, compared to 40 minutes before, caused a significantly greater delay the timing to the first hypoglycemia episode (P=.02). More participants required carbohydrate treatment for hypoglycemia during exercise under the T-40 condition than under the T-90 condition, although this was not significantly different (P=.15). During the preexercise period, free insulin concentrations tended to decrease more under the T-90 condition than under the T-40 condition (P=.08). Local muscle vasodilation was similar between basal insulin reduction strategies; however, in instances with higher exercise-induced muscle vasodilation, there was also a greater drop in glucose concentrations (P<.005).
Conclusion
When participants performed 60 minutes of moderate-intensity exercise, plasma glucose decreased less if an 80% basal insulin reduction was set 90 minutes before exercise onset than when it was set 40 minutes before exercise onset. T-90 delayed the onset of hypoglycemia during exercise significantly more than T-40; however, there was no significant difference in the overall number of hypoglycemia events between conditions.
No More Hypoglycaemia on Days with Physical Activity and Unrestricted Diet when Using a Closed-Loop System for 12 Weeks: a Post Hoc Secondary Analysis of the Multicentre, Randomized Controlled Diabeloop WP7 Trial
Franc S1,2,3, Benhamou PY4, Borot S5, Chaillous L6, Delemer B7, Doron M8, Guerci B9, Hanaire H10, Huneker E11, Jeandidier N12, Amadou C1,13, Renard E14, Reznik Y15, Schaepelynck P16, Simon C17, Thivolet C18, Thomas C3, Hannaert P19, Charpentier G2,3
1Department of Diabetes, Sud-Francilien Hospital, Corbeil-Essonnes, France; 2Centre d'Etude et de Recherche pour l'Intensification du Traitement du Diabète (CERITD), Evry, France; 3Laboratoire de Biologie de l'Exercice pour la Performance et la Santé, Université Evry Val d'Essonne, Institut de Recherches Biomédicales des Armées, Université Paris Saclay, Evry, France; 4Department of Diabetology, University Hospital Grenoble Alpes, Grenoble, France; 5Department of Endocrinology, Metabolism, Diabetes and Nutrition, Centre Hospitalier Universitaire Jean Minjoz, Besancon, France; 6CHU de Nantes - Hospital Laennec, Saint-Herblain, France; 7Department of Endocrinology, Diabetes and Nutrition, Reims University Hospital, Reims, France; 8Université Grenoble Alpes, Commissariat à l'Energie Atomique, Laboratoire d'électronique et de technologie de l'information, Département micro Technologies pour la Biologie et la Santé, Grenoble, France; 9Endocrinology-Diabetes Care Unit, University of Lorraine, Vandoeuvre Lès Nancy, France; 10Department of Diabetology, Metabolic Diseases and Nutrition, CHU Toulouse, University of Toulouse, Toulouse, France; 11Diabeloop S.A., Paris, France; 12Department of Endocrinology, Diabetes and Nutrition, CHRU of Strasbourg (UDS), Strasbourg, France; 13University Paris-Saclay, Orsay, France; 14Department of Endocrinology, Diabetes and Nutrition, Montpellier University Hospital, and Institute of Functional Genomics, CNRS, INSERM, University of Montpellier, Montpellier, France; 15Department of Endocrinology, University of Caen Côte de Nacre Regional Hospital Center, Caen, France; 16Department of Nutrition-Endocrinology-Metabolic Disorders, Marseille University Hospital, Sainte Marguerite Hospital, Marseille, France; 17Department of Endocrinology, Diabetes and Nutrition, Centre Hospitalier Lyon Sud, Lyon, France; 18Center for Diabetes, Lyon University Hospital, Lyon, France; 19School of Medicine and Pharmacy of Poitiers, Ischémie Reperfusion en Transplantation d'Organes Mécanismes et Innovations Thérapeutiques, Institut National de la Santé et de la Recherche Médicale, Unité Mixte de Recherche 1082, Poitiers, France
Background
Automated insulin delivery (AID) and closed-loop technology increase glucose time-in-range (TIR) more than conventional pump therapy. However, AID systems often do not fully protect against the rapid decline in glycemia commonly associated with moderate-intensity aerobic exercise. This post-hoc analysis investigated the efficacy of the Diabeloop Generation-1 (DBLG1) AID system managing the increased risk of hypoglycemia associated with physical activity.
Methods
The Diabeloop WP7 study was a 12-week, multicenter, open-label, randomized controlled crossover trial of DBLG1 in 68 adults with type 1 diabetes (T1D) in free-living conditions. The primary outcome was the difference in time below range (TBR) between days with physical activity (PA) compared with those without PA. Secondary outcomes included timing of activity announcement, preventative carbohydrates, insulin delivery, mean glucose, TIR, and time above range (TAR).
Results
Of the 68 participants that completed the 12-week trial, seven participants did not record any PA and were excluded from the analysis. A total of 56 adults with T1D recorded 19.9±24.1 PA events/participant. The duration of PA was 82.3±58.3 minutes with 40% performing mild PA, 41% moderate PA, and 19% intense PA. No significant difference in TBR (P=.282) was seen between days with PA and those without. Similarly, overnight TBR was not different on days after PA than on days after no PA (P=.507). Preventative carbohydrate intake suggested by DBLG1 and recorded carbohydrate intake was significantly higher with PA than without PA (P<.0001).
Conclusion
In free-living conditions, adults with T1D using DBLG1 experienced no more hypoglycemia on days with PA than on days without activity, independent of intensity and duration of activity. In addition, carbohydrate intake recommendations by the system and TAR were significantly higher on days with activity.
Comments
If you were to think about hypoglycemia prevention strategies for exercise in individuals with T1D, what typically comes to mind first? For us, it is usually 1) consuming carbohydrates and/or 2) insulin adjustments well in advance of exercise (if we remember!). As many studies in this article have shown, announcing exercise, setting higher glucose targets at least 1 to 2 hours before exercise, and/or consuming carbohydrates before or during exercise seem to be repeating strategies.
The Tagougui et al. paper demonstrates that an 80% basal rate insulin reduction is more effective at delaying hypoglycemia when set 90 minutes before than when set 40 minutes before 1 hour of moderate-intensity exercise. Since there were no significant differences in the absolute number of hypoglycemia events in this study, additional carbohydrate consumption with a basal rate insulin reduction may be just what is needed for some individuals who forget to set the reduced basal rate well in advance of activity.
In the study by Franc et al., the researchers found that while participants were using the Diabeloop Generation-1 (DBLG1) system, their TBR (<70 mg/dL) was not significantly affected by exercise, irrespective of intensity or duration of the activity. However, this outcome was largely accompanied by an increase in carbohydrate intake and a reduction in insulin delivery on exercise days. In fact, when “physical activity mode” is entered at least 2 hours before exercise in the DBLG1 system, and if glycemia forecast is less than 160 mg/dL at exercise onset, depending on the intensity and deviation from the glucose target, a carbohydrate snack is recommended to further protect against hypoglycemia. Interestingly, participants in this study consumed around 30% more than that recommended by the DBLG1 system, which suggests they were still worried about exercise-associated hypoglycemia despite being on an AID system. So perhaps more work is needed if we want to eliminate or significantly reduce carbohydrate snacking for exercise in T1D!
Footnotes
Author Disclosure Statement
Michael C. Riddell has received, in the last 12 months, speaker fees from Novo Nordisk, Dexcom and Sanofi and has acted as a consultant and/or an advisor for Supersapiens, Indigo Diabetes, Zealand Pharma, Zucara Therapeutics, and the JAEB Center for Health Research. Dessi P. Zaharieva has received speaking honoraria from Medtronic Diabetes, Ascensia Diabetes, Insulet, and the American Diabetes Association.