Exercise-Induced Hypoglycemia in IDDM Patients Treated with a Short-Acting Insulin Analogue

J.A. Tuomine, S.L. Karonen, L. Melamies,
G. Bolli, V.A. Koivisto
Diabetologia 38:106-11, 1995

Summary and commentary by
Patrick J. Boyle, MD

Objective. To demonstrate the effect of the insulin analog lispro on exercise-induced hypoglycemia related to timing of dose in patients with type I diabetes.

Design. All patients cycled on a stationary bike on four different mornings. On two of these mornings, exercise occurred 40–80 minutes after the beginning of a 550 kcal breakfast (early exercise). On the other two mornings, exercise occurred 180–220 minutes after breakfast (late exercise). Before one of the early- and one of the late-exercise mornings, an average of 6.3 U of regular human insulin was injected subcutaneously 30 minutes before the meal. An identical amount of lispro was injected 5 minutes before breakfast on the other two days. An additional 30-gram carbohydrate snack was provided at 120 minutes on each of the mornings.

Subjects. Ten healthy patients with type I diabetes (seven male, three female) with an average duration of diabetes of 18 years and taking an average insulin dose of 49 U/day agreed to participate. All patients were using regular insulin before each meal and NPH insulin at bedtime. In addition, three of the patients were also using NPH insulin in the morning with regular insulin before breakfast.

Measurements. Plasma glucose concentration was determined at 10- and 20-minute intervals during the 240 minutes after ingestion of breakfast. In addition, serum insulin, glucagon, cortisol, and catecholamines were determined frequently throughout the study period.

Results. Glucose concentrations fell during exercise in all four conditions. During the early exercise periods, the mean glucose concentration fell by ~80 mg/dl (~4.5 mmol/L) in association with lispro administration. A fall of 36 mg/dl followed injection of regular insulin (P < 0.01). One patient in each treatment group required supplemental carbohydrate before exercise due to plasma glucose concentration that fell below 54 mg/dl.

During late exercise, mean average glucose concentrations fell by only 18 mg/dl following lispro injection, but

by nearly 36 mg/dl in the regular insulin-treated group (P < 0.05). As glucose concentration fell with early exercise, glucagon concentrations were higher after lispro than after regular insulin use.

Conclusions. The lispro analog of human insulin can either augment or reduce exercise-induced hypoglycemia, depending on the time interval between insulin injection and exercise.

Commentary
Exercise is fundamental in treating patients with either type I or insulin-requiring type II diabetes mellitus. The enhanced insulin sensitivity that is seen following the exercise period translates into lower exogenous insulin requirements.

A frequently overlooked long-term benefit is the associated increase in energy expenditure. Intensively managed patients are at greater risk of gaining weight.¹ Therefore, treatment strategies that incorporate exercise would be expected to minimize weight gain in association with better overall glucose control.

The cardiovascular and lipid-lowering benefits of exercise are well known, and, given the three- to fivefold greater likelihood of macrovascular disease in patients with diabetes compared to people without diabetes, exercise is key in reducing morbidity and mortality in the long run of their lives.

Given these obvious advantages, the dilemma arises as to how health-care providers can most safely prescribe exercise programs for patients with insulin-requiring diabetes. Unfortunately, exercise still needs to be a planned activity coordinated with the time of the last insulin injection and food intake to minimize hypoglycemia.

Teaching patients to take advantage of the absorption kinetics of the insulin they use is essential in minimizing hypoglycemia. The study by Tuomine and associates demonstrates that exercise occurring 3 hours after a morning meal in conjunction with lispro administration is less likely to be associated with a major drop in glucose concentration than that expected to occur late in the peak of action of conventional regular insulin. Patients have a variety of preferences and needs as to when they will be physically active or have time to exercise. Thus the advent of a more rapidly acting insulin will presumably make exercise several hours after a meal less likely to be associated with hypoglycemia than current preparations of regular insulin are.

Several observations are warranted in the study under review. Both mornings of early exercise started with what would be considered unacceptable metabolic control-fasting glucose concentrations were approximately 200 mg/dl. In general, exercising in the face of hyperglycemia and waning insulin concentrations enhances rates of glucose production by the end of the exercise period.² This phenomenon is most likely linked to epinephrine release during exercise, which promotes hepatic glucogenesis and glycogenolysis coupled with poor glucose disposal. Therefore, the days of early exercise depict what might occur in less-than-optimal metabolic control, and extrapolation to what might occur when glucose concentrations are nearer to the desirable, normal range should be done with some care.

Given the rather extraordinary exercise-related fall in glucose concentration seen coincident with the higher peak insulin concentrations following lispro administration, one might expect a much greater frequency of hypoglycemia when patients with normal preprandial glucose concentrations exercise shortly after a meal that includes lispro injection. In this sense, if one plans exercise to happen shortly after breakfast and has gone into that meal with a normal glucose concentration, using regular human insulin may actually be to the patient’s advantage. Patients who use lispro exclusively may need to reduce their dose and/or include more rapidly absorbable carbohydrate in their plan.

At the beginning of the late exercise periods, the glucose concentration of patients having received regular insulin was near 210 mg/dl after their mid-morning snack. Had these patients gone into the late exercise period with a more normal glucose concentration, they would undoubtedly have experienced substantially more episodes of hypoglycemia than the authors observed. The major lesson suggested, but not proven, by this study is that exercise conducted during the peak of insulin action in association with near-normal pre-exercise glucose concentrations will likely lead to hypoglycemia.

The location of the injection is not trivial when developing an exercise routine. Subcutaneous injection of insulin in the thigh leads to more rapid absorption during exercise than injection into abdominal fat.³ Patients must understand that exercise increases blood flow to the muscle groups doing the work and that the heat produced from the work is partially dissipated by increasing blood flow to the skin overlying these muscle groups. The side effect is that rates of insulin absorption in the extremity are accelerated during and after exercise. Therefore, cycling, as done in this experiment, is least likely to be associated with hypoglycemia when insulin is injected into the abdomen as compared with the leg.

Given the pivotal role that exercise plays in the lives of the patients for whom we care, specific exercise plans that match patients’ preferences with their cardiovascular fitness, as well as take into consideration their specific insulin therapy, are integral to long-term acceptance and adherence to exercise. Lispro has a multitude of significant lifestyle advantages in terms of flexibility in dosing in association with food consumption. Tuomine and associates’ investigation should prompt health-care providers switching patients to lispro to spend a few moments going over when major decrements in glucose concentration are expected to occur after injection and when to exercise to minimize hypoglycemia.

References

¹The DCCT Research Group: Weight gain associated with intensive therapy in the Diabetes Control and Complications Trial. Diabetes Care 11:567-73, 1988.

²Berger M, Berchtold P, Cuppers HJ, et al: Metabolic and hormonal effects of muscular exercise in juvenile diabetics. Diabetologia 13:355-65, 1977.

³Koivisto VA, Felig P: Effects of leg exercise on insulin absorption in diabetic patients. N Engl J Med 298:79-83, 1979.

Patrick J. Boyle, MD, is an associate professor in the Division of Endo-crinology, Diabetes, and Metabolism and the director of diabetes case management at the University of New Mexico Health Sciences Center at Albuquerque.

Note of disclosure: Dr. Boyle has received honoraria for speaking engagements from Eli Lilly and Company, which manufactures lispro and other products for the treatment of diabetes.