Exercise and glycemic imbalances: A situation-specific estimate of glucose supplementation
Journal Title (Medline/Pubmed accepted abbreviation):  Med. Sci. Sports Exerc.
Year: 2011
Volume: 43
Number: 1
Page numbers: 2-11
doi (if applicable): 10.1249/MSS.0b013e3181e6d6a1

Summary of Background and Research Design

Background:Insulin is a hormone that stimulates body cells, including muscle, to take up glucose from the blood stream. Patients with type 1 diabetes do not produce insulin and therefore must administer it themselves in conjunction with a meal. During exercise, muscle takes up glucose at a faster rate than at rest. In healthy subjects, the liver replenishes blood glucose via tapping into glycogen stores or gluconeogenesis. However, patients with type 1 diabetes often do not replenish glucose at an adequate rate and they experience hypoglycemia. To circumvent hypoglycemia, lower injections of insulin at the pre-workout meal have been proposed. However, this necessitates the anticipation of exercise. Ingesting additional carbohydrate is another possible solution. The patient’s physical fitness, exercise intensity, exercise duration, current insulin levels (effected by time of day), and current glucose levels are all factors in the calculation of how to control metabolism, and it often requires several attempts for the individual to learn what his or her body requires to prevent hypoglycemia. This can be very discouraging and/or dangerous.

Research Aim:An algorithm called “Exercise Carbohydrate Requirement Estimating Software” (ECRES) was created in which a patient can enter personal and situational information in order to calculate how much carbohydrate should be ingested before and during/after moderate exercise to prevent hypo- (and hyper-) glycemia. This article explains the program in detail and evaluates its accuracy.

The algorithm:
Input: Pharmacokinetic profiles of insulin and other literature values, the patient’s insulin levels throughout a 24 hr period with their usual insulin therapy (empirically determined), normal carbohydrate intake and timing, normal training habits, exercise specific data including intensity and duration, daytime schedule, and starting glycemia.
Output: Amount of carbohydrates that should be consumed before/during the exercise and after the exercise
The software is capable of accounting for any reduction in insulin administration during a pre-workout meal.

Testing the algorithm:
Subjects: Twenty-seven patients with type 1 diabetes (19 men and 8 women, age 44 ± 11, diagnosed 22 ± 11 yrs before recruitment). Subjects were categorized as “aerobically trained”, who regularly exercised at least 30 min 3 days/wk, or “sedentary”.

Treatments and protocol:Subjects were encouraged to follow normal practices, including their normal procedures for avoiding exercise-induced hypoglycemic events. Participants completed 3x 1 hr treadmill walks starting 90, 180, or 270 min after lunch. Each participant completed all three time periods in a random order. Speed and/or incline were adjusted to maintain intensity of 65% of their maximum heart rate (220 – age). The participants’ normal diet and insulin dosing was followed except for the 90 min trial. For these, participants were asked to consume half of their normal amount of carbohydrates and inject half the amount of insulin. In total, 81trials were performed (27 subjects x 3 trials/subject).
                Participants arrived at the laboratory 45 min before the walk. Glycemia and blood concentrations of ketone bodies were measured and the software was run. Patients were given 70% of the amount of carbohydrates estimated by ECRES to counter hypoglycemia during the walk in the form of sugar or sugar drops. Glycemia was measured at the start of the trial and every 15 min until the end of the walk. The remaining 30% of carbohydrates were given during exercise, part after 15 min elapsed and the other part after 30 min. If glycemia dropped below 5.0 mM, patients were given 5 g additional sugar. Late-onset hypoglycemia over the following 24 hrs was avoided with a special protocol.

Summary of research findings:
Altogether, ECRES succeeding in maintaining blood glucose between 4.0 mM and 10.0 mM for 57 of the 81 trials. ECRES calculated an insufficient amount of glucose for 11 of 81 walks and excess glucose for 13 trials. The unsatisfactory calculations did not correlate with the time of day that the exercise occurred. In most trials, the amount of energy ingested was smaller than the amount utilized during exercise. Therefore, exercise can be performed safely in weight control programs.
                The largest sources of error in the calculation were: 1) the assignment of a participant as “trained” or “sedentary”. This distinction is made because the amount of glucose oxidized during exercise by a sedentary individual can be notably higher than that of a trained individual. This dichotomy is not clear cut and can lead to errors. 2) Individuals’ insulin sensitivity was calculated by the algorithm from the ratio between the amount of dietary carbohydrates and the patient’s usual insulin dose but proved to be more complicated than that. More optimization is needed. 3) Hypoglycemic events can occur during normal activity for unknown reasons that may or may not be able to be accounted for in future versions of the program.

Interpretation of findings/Key practice applications:

The ECRES algorithm was successful in estimating the amount of carbohydrate that should be ingested to avoid a hypoglycemic event during exercise in 70.4% of the trials. It is a good application for persons with type 1 diabetes who may need to calculate how much carbohydrate to ingest before unplanned activity. Further optimization of the algorithm including a more accurate assessment of the individual’s fitness/activity level is needed for more useful results.


This algorithm, thus far, is only applicable to moderate exercise (55-70% of maximal heart rate) and therefore cannot be applied to more intense physical activity. Moreover, it was only tested for exactly 60 min of exercise. Also, sugar was the only source of carbohydrate (the identity of the sugar was not stated- sucrose? glucose?). Other carbohydrate sources and foods eaten concomitantly can have very different glycemic responses.
Google Tracking Google Plus Tracking Twitter Tracking