Original Articles

c omparison of the Acute and Long'Term

Effects of Exercise on Glucose Control
in Type I Diabetes
B. ZINMAN, M.D., S. ZUNIGA-GUAJARDO, M.D., AND D. KELLY, M.D.
To determine the long-term effect of exercise training on glucose control, 13 subjects with type I
diabetes and 7 control subjects performed 45 min of cycle exercise three times per wk for 12 wk. The
acute blood glucose response, the long-term effect on glucose control (glycosylated hemoglobin and
fasting plasma glucose), and changes in nutrient intake were assessed. Fitness as measured by VO2 MAX
increased in both control (33.8 ± 1.7 to 43.2 ± 3.5 ml/min/kg) and diabetic (38.7 ± 3.3 to 46.5 ± 3.6
ml/min/kg) (P < 0.05) subjects although body weight remained unchanged. In the diabetic subjects,
an acute glucose-lowering effect occurred with each exercise session throughout the 12-wk training
period (225.8 ± 16.1 to 148.5 ± 16.8 mg/dl, P < 0.001). However, fasting plasma glucose and gly-
cosylated hemoglobin remained essentially unchanged (pretraining, 193.7 ± 27.5 mg/dl and 10.7 ±
0.3%; 6-wk training, 192.5 ± 27.1 mg/dl and 10.7 ± 0.03%; 12-wk training, 202 ± 30.1 mg/dl and
10.3 ± 0.8%). Total caloric intake as assessed by diet history increased significantly on exercising days
(2569-2849 kcal, P < 0.05). Although plasma glucose decreases acutely with exercise, increased caloric
intake on exercising days obviates a long-term effect of training on glucose control. More precise
guidelines and recommendations as to exercise timing and nutrient intake, likely based on self-moni-
toring of blood glucose, are required to achieve a beneficial effect of exercise training on metabolic
control in type I diabetes, DIABETES CARE 1984; 7:515-19.

S
ince the early demonstration of its acute glucose-
lowering effect,1 exercise has been considered an
important component of the management of dia-
betes.2 With the increased interest in fitness and
exercise in general, the role of exercise in the treatment of
diabetes has recently been re-emphasized. However, docu-
mentation of a beneficial long-term effect on glucose control
in type I diabetes has been lacking.
Most recent studies have focused on defining the metabolic
responses to acute exercise in type I diabetes as compared
with normal controls. 3~7 The effects of insulin deficiency and
insulin excess on glucoregulation and the mobilization of
energy stores for exercising muscle have been well defined.
As a result of these studies, specific guidelines on the avoid-
ance of exercise-induced hypoglycemia have been suggested.
In addition, it has been shown that with intravenous insulin
infusion pumps specific insulin infusion waveforms (algo-
rithms) for exercise both in the fasted and postprandial states
could be developed, which normalize the metabolic responses
to exercise.8'9 However, these studies do not address the issue
of the therapeutic benefits of exercise on glucoregulation in
diabetes. Limited investigations in this regard have shown
that physical training increases insulin sensitivity in obese
subjects and patients with type II diabetes, although a long-
term effect on improving glucose tolerance was not consis-
tently evident. 10~13 A recent study examining the response to
training of patients with type I diabetes snowed increased
peripheral insulin sensitivity but unchanged blood glucose
control.14 This later study appears to be inconsistent with
the well-documented hypoglycemic effects of acute exercise
and the observation that postprandial exercise significantly
improves meal-related glucose excursions.15 One would ex-
pect these acute responses to have a summation effect on
overall glucose control in the training patient.
To date, few studies have examined both the acute and
long-term effects of exercise in the same population. The
present study was thus undertaken to document both the
acute and long-term effects of 45 min of exercise performed
three times per wk in type I diabetic individuals as compared
with normal controls. Measurements of blood glucose were

DIABETES CARE, VOL. 7 NO. 6, NOVEMBER-DECEMBER 1984 515

 EFFECTS OF EXERCISE ON GLUCOSE CONTROL IN TYPE 1 DIABETES/B. ZINMAN AND ASSOCIATES
mg/dl
250
• Bete>re exercise * p <0.01
O AMe r exercise ** p < 0.05
200 -
I T I at two different workloads measuring oxygen uptake and heart
*
ill * rapid response CO2 (LB/2) andO 2 (OM/11) analyzers (Beck-
150 -
I I PI
I man Instruments Inc., Palo Alto, California). Once VO2
mm
100
week 0 to 6
S§j
week 6 to 12
mm
week 0 to 12
FIG. I. Blood glucose concentrations before {hatched bars) and after
(open bars) each exercise session during the first 6 wk of training, second
6 wk of training, and over the entire training period. Mean ± SEM are
shown for 13 subjects with type I diabetes.
made before and after each exercise session while fitness,
glycosylated hemoglobin, and fasting plasma glucose were
assessed before and after 6 and 12 wk of exercise training.
MATERIALS AND METHODS
Subjects. Thirteen type I diabetic subjects and seven normal
controls were studied. All were nonobese. The six female
and seven male diabetic subjects were 30.0 ± 1.8 yr of age,
97.5% of ideal body weight, with a duration of diabetes of
14.3 ± 3.2 yr and an average daily insulin dose of 37.6 ± 3.2
U. Eleven of the diabetic subjects were taking a combination
of intermediate-acting insulin (NPH or lente) and regular
(crystalline zinc) insulin, and two were being treated with
intermediate-acting insulin alone. Ten of the diabetic sub-
jects were receiving one insulin injection daily while two
were on two injections and one was on three injections. The
control group consisted of five women and two men with a
mean age of 30.7 ± 2.9 yr. They were 101 ± 5% of ideal
body weight and had no family history of diabetes mellitus,
nor were they taking any medication. None of the diabetic
patients had clinical evidence of diabetic microvascular or
macrovascular complications, although two were receiving
thyroid replacement for primary hypothyroidism. All subjects
had a normal exercise electrocardiogram and informed con-
sent was obtained in compliance with the Human Subjects
Review Committee of the University of Toronto.
TABLE 1
Effect of exercise training on maximal oxygen uptake (V0 2 MAX) (ml/min/kg)
Pretraining
(Owk)
Control subjects 33.8 ± 1.7 40.0 ± 1.8*
Diabetic subjects 38.7 ± 3.3 44.6 ± 3.0'
'P < 0.05 as compared with pretraining.
516 DIABETES CARE, VOL. 7 NO. 6, NOVEMBER-DECEMBER 1984
Exercise training program. Maximal oxygen uptake (VO 2
MAX) was estimated during steady-state exercise on a bicycle
ergometer (Siemens-Elema Model 380, Stockholm, Sweden)
rate.1617 Respiratory gas concentrations were measured with
MAX was assessed and baseline metabolic measurements per-
formed, physical training was initiated 3 times per wk for 12
wk under the supervision of a qualified exercise therapist.
Training consisted of 45 min of exercise on a stationary cycle
ergometer with a workload of sufficient intensity to stimulate
an increase in endurance fitness.l8 Heart rate response achieved
during each exercise session was 60-85% of predicted max-
imum. The actual work intensity was individually determined
and adjusted by the results obtained from the fitness assess-
ment. Training workloads were reassessed every 2 wk and
appropriate changes made to result in continued stimulation
and cardiorespiratory adaptation. Before and after each ex-
ercise session, fingerprick blood samples were taken for glu-
cose measurement using glucose-oxidase reagent strips and a
glucometer (Ames, Miles Laboratories, Elkhart, Indiana).
Blood samples were taken before training, at 6 and 12 wk of
training, and in the posttraining period for determination of
glycosylated hemoglobin and fasting plasma glucose. Ten of
the 13 diabetic patients were interviewed by a qualified die-
titian after the training period in order to assess caloric intake
on exercising and nonexercising days by diet history. Diet
histories were hand-calculated from reference guides (Nu-
trient Values of Some Common Foods, Health and Welfare
Canada, 1979, and the National Heart, Lung and Blood
Institute Nutrition Coding Centre Food Table Intake,
1977). 19~21 In two patients, caloric intake was assessed from
5-day food records.
Subjects were asked to continue with their daily routines
and usual diets. They were instructed to record all hypo-
glycemic episodes and any adjustments in insulin dosage.
They were asked to compensate for exercise as they were
usually instructed to by their physician.
All subjects administered their insulin 15-30 min before
breakfast. Five subjects performed the exercise training 30
min after breakfast (0800-0830 h), four subjects exercised
90-120 min after breakfast (0930-1000 h), and four subjects
exercised 30 min after lunch (1230-1300 h). There was no
difference in the observed response in relationship to the
time of exercise.
Training
Posttraining
6 wk 12 wk (18 wk)
43.2 ±3.5* 40.0 ± 4.0
46.5 ±3.6' 43.1 ± 3.3*
 EFFECTS OF EXERCISE ON GLUCOSE CONTROL IN TYPE I DIABETES/B. ZINMAN AND ASSOCIATES

TABLE 2
Effect of exercise training on weight (kg)

Training
Pretraining Posttraining
(0 wk) 6 wk 12 wk (18 wk)

Control subjects 65.7 ± 5.2 65.4 ± 5.1 65.1 ± 5. 2 64.9 ± 5.1

Diabetic subjects 64.2 ± 1.1 64.3 ± 1.1 64.2 ± 1. 1 64.5 ± 1.1

Analytic methods. All venous blood samples were obtained values. Body weight was similar in both control and diabetic
in the postabsorptive state after a 10-h fast. Plasma glucose subjects at the start of training and no significant changes
was determined using a glucose analyzer (Beckman Instru- were seen in either group (Table 2).
ments, Inc.) and glycosylated hemoglobin concentrations To determine the effect of exercise on overall glucose con-
(hemoglobin Al) were determined by the microcolumn tech- trol, fasting plasma glucose and hemoglobin Al measure-
nique with preincubation of samples to remove the reversible ments were performed before, during, and after training (Ta-
glycosylated fractions. To compare differences within groups, ble 3). The diabetic subjects had elevated fasting plasma
Student's paired t-test was used, and for differences between glucose (193.7 mg/dl) and hemoglobin Al (10.7%) (P <
groups Student's unpaired t-test was employed. All data are 0.005) as compared with control subjects before training and
given as the mean ± SEM in the tables and figures. neither changed significantly with training. Other clinical
measures of diabetes control were also recorded before and
RESULTS after training. Total insulin dose was unchanged at 37.6 ± 3.2
U before and 40.0 ± 3.1 U after training, and the occurrence
Acute response to exercise. The heart rate response during the of hypoglycemic reactions was similar.
last 10 min of each training session was similar for both The effect of exercise on nutrient intake as assessed by
control and diabetic subjects (156.7 ± 1.9 beats/min and diet history is shown in Table 4. Total caloric intake in-
160 ± 3.2 beats/min). Figure 1 shows the acute blood glu- creased significantly on exercising days and this was primarily
cose response to each exercise session during the first 6 wk in the form of carbohydrate and protein. Two subjects kept
of training (first pair of bars), the second 6 wk of exercise 5-day food records that showed a similar increase in ingested
training (middle bars), and over the entire training period calories on exercising days (520 and 828 kcal, respectively).
(third pair of bars). A significant fall in glycemia occurred
with each exercise session and this acute decrease was similar
DISCUSSION
and persisted throughout the entire exercise training period.

A
Effects of exercise training. Maximal oxygen uptake (VO2 lthough exercise is thought to be useful in the
MAX) is shown in Table 1. VO2 MAX increased 28% and treatment of diabetes, there is little evidence to
20% in the control and diabetic groups, respectively, at 12 support a beneficial effect of exercise training in
wk of training. When reassessed at 18 wk, VO2 MAX de- type I diabetes. Indeed, a recent Scandinavian
creased in both groups. However, a significant difference study14 failed to demonstrate a significant change in glycos-
persisted in the diabetic subjects as compared with pretraining ylated hemoglobin in 9 type I diabetic subjects undergoing

TABLE 3
Effect of exercise training on fasting plasma glucose and glycosylated hemoglobin

Training
Pretraining Posttraining
(Owk) 6 wk 12 wk (18 wk)

Fasting plasma glucose,

mg/dl (mmol/L)
Control subjects 87.4 ± 3.6 87.2 ± 6.3 84.6 ± 2.7 81.0 ± 2.5
(4.8 ± 0.2) (4.8 ± 0.4) (4.7 ± 0.2) (4.5 ±0.1)
Diabetic subjects 193.7 ± 27.7* 192.5 ± 27.1* 202 ± 30. T 189.8 ± 29.5*
(10.8 ± 1.5) (10.7 ± 1.5) (11.2 ± 1.7) (10.5 ± 1.6)
Hemoglobin Al (%)t
Control subjects 7.7 ± 0 . 5 7.5 ± 0.4 7.4 ± 0.2 7.0 ± 0.1
Diabetic subjects 10.7 ± 0.3* 10.7 ± 0.3* 10.3 ± 0.8' 10.6 ± 0.3*

'P < 0.005 as compared with control subjects.

tNormal, 6.5-8.5%.

DIABETES CARE, VOL. 7 NO. 6, NOVEMBER-DECEMBER 1984 517

 EFFECTS OF EXERCISE ON GLUCOSE CONTROL IN TYPE I DIABETES/B. Z1NMAN AND ASSOCIATES
TABLE 4
Effect of exercise on nutrient intake
Nonexercising day Exercising day
Total calories (kcal) 2569 ± 273 2849
Carbohydrate (g) 265 ± 30 309
Protein (g) 115 ± 11 125
Fat (g) 115 ± 15 121
'P < 0.05 as compared with nonexercising day.
16 wk of exercise training despite an increase in peripheral
insulin sensitivity. However, the patients studied tended to
be physically fit before training (3.24 ± 0.16 L/min or 42
ml/min/kg), and only an 8% increase in VO2 MAX was achieved
with training.
The failure to demonstrate a beneficial effect of exercise
on long-term glucose control appears to be inconsistent with
the well-known acute hypoglycemic effect of exercise1"7 and
the observation that exercise can improve postprandial glu-
cose excursion.15 The present study was undertaken to ex-
amine the variables that might explain this apparent con-
tradiction. Thus, the acute glucose responses to exercise, the
effect of exercise on nutrient intake, and the long-term re-
sponses to exercise training were studied in 13 type I diabetic
subjects.
As shown in Table I, a significant increase in fitness as
measured by VO2 MAX was achieved in both control (33.8 ±
1.7 to 43.2 ± 3.5 ml/min/kg) and diabetic (38.7 ± 3.3 to
46.5 ± 3.6 ml/min/kg) subjects with 12 wk of exercise train-
ing on a cycle ergometer. The acute glucose-lowering effect
of exercise was present with each exercise session and the
magnitude of this effect extended over the entire training
period (Figure 1). Since exercise was performed at least 3
times per wk, one would expect that the repeated acute
decreases in blood glucose with exercise would be reflected
in a long-term decrease in glycosylated hemoglobin. How-
ever, this was not the case, and hemoglobin Al remained
constant throughout. Since the variable of nutrient intake
clearly plays an important role in overall diabetes control,
diet histories were obtained by a qualified dietitian to assess
caloric intake on exercising and nonexercising days. As
shown in Table 4, a significant increase in caloric intake
was evident during exercising days. A similar increase was
seen in two patients who kept 5-day food records. These
findings are consistent with the common clinical practice of
avoiding exercise-induced hypoglycemia by ingesting supple-
mental calories. However, without specific guidelines that
could be developed from self-monitoring of blood glucose,
overtreatment commonly occurs. In addition, several pa-
tients clearly indicated that during exercise days they could
"liberalize" their diet and considered this one of the advan-
tages of regular exercise. The failure to demonstrate a sig-
nificant difference in body weight with training is also con-
sistent with the observations of increased food consumption
on exercising days. Interestingly, this apparent self-regulation
518 DIABETES CARE, VOL. 7 NO. 6, NOVEMBER-DECEMBER 1984
of caloric intake to match increased caloric expenditure with
the maintenance of body weight has been previously docu-
mented experimentally in rodents.22 Although changes in
body composition were not precisely assessed in our study,
± 330' triceps skinfold thickness did not change with training (con-
± 38* trol subjects: 14.3 ± 2.0 versus 14.0 ± 2.3 mm; diabetic
± 15* subjects: 15.3 ± 2.3 versus 14-2 ± 2.3 mm, before and after
± 17 training, respectively).
In summary, a significant increase in fitness as assessed by
VO2 MAX can be achieved with 12 wk of exercise training
on a cycle ergometer in type I diabetes. Although plasma
glucose decreased acutely with each exercise session, a long-
term effect on overall glucose control could not be demon-
strated. Increased caloric intake primarily as carbohydrate on
exercising days likely accounts for the above observations.
Although exercise training per se does not have a significant
effect on overall glucose control, diabetic patients should be
encouraged and have the freedom to exercise for the same
general reasons (primarily cardiovascular and psychological)
that it is recommended for the population at large.23"27 To
demonstrate a beneficial metabolic effect of exercise training
on diabetes control, more precise guidelines and recommen-
dations as to exercise timing and nutrient intake based on
self-monitoring of blood glucose data may be required.
ACKNOWLEDGMENTS: We are indebted to Patricia A. McClean
and other members of the respiratory research laboratory for
technical support; to Kathy Camelon and Ann Murphy for
the dietary evaluation; to the Tri-Hospital Diabetes Educa-
tion Centre for administrative support; and to Mae Abucay
for excellent secretarial assistance in preparation of the man-
uscript.
These studies were supported by a grant from the Canadian
Diabetes Association and the Banting and Best Diabetes
Centre.
This paper was presented in part in abstract form at the Ca-
nadian Society for Clinical Investigation Meeting in September
1983 (Clin. Invest. Med. 1983; 6:69).
From the Division of Endocrinology and Metabolism, Toronto
General Hospital, Department of Medicine, University of To-
ronto, Toronto, Ontario, Canada.
Address reprint requests to Dr. B. Zinman, Toronto General
Hospital, 101 College Street, Toronto, Ontario M5G 1L7, Can-
ada.
REFERENCES
1
Lawrence, R. D.: The effect of exercise on insulin action in
diabetes. Br. Med. J. 1926; 1:648-52.
2
Marble, A., and Smith, R. M.: Exercise in diabetes mellitus.
Arch. Intern. Med. 1936; 58:577-88.
3
Vranic, M., and Berger, M.: Exercise in diabetes mellitus. Di-
abetes 1979; 28:147-67.
4
Felig, P., and Wahren, J.: Fuel homeostasis in exercise. N. Engl.
J. Med. 1975; 293:1978-84.
5
Zinman, B., Vranic, M., Albisser, A. M., and Marliss, E. B.:
The role of insulin in metabolic response to exercise in diabetic
man. In Proceedings of the KROC Conference on Exercise and
 EFFECTS OF EXERCISE ON GLUCOSE CONTROL IN TYPE I DIABETES/B. ZINMAN AND ASSOCIATES
Diabetes. Vranic, M., Wahren, J., and Horvath, S., Eds. Diabetes
1979; 28 (Suppl. 1):76-81.
6
Zinman, B., Murray, F. T., Vranic, M., Albisser, A. M., and
Marliss, E. B.: Glucoregulation during moderate exercise in insulin
treated diabetes. J. Clin. Endocrinol. Metab. 1977; 45:641-52.
7
Murray, F. T., Zinman, B., McClean, P. A., Devoga, A.,
Albisser, A. M., Leibel, B. S., Nakhooda, A. F., Stokes, E. F,
and Marliss, E. B.: The metabolic response to moderate exercise
in diabetic man receiving intravenous and subcutaneous insulin. J.
Clin. Endocrinol. Metab. 1977; 44:708-20.
8 19
Nelson, J. D., Poussier, P., Marliss, E. B., Albisser, A. M.,
and Zinman, B.: The metabolic responses of normal and insulin
infused diabetics to postprandial exercise. Am. J. Physiol. 1982;
242:E309.
9 20
Poussier, P., Zinman, B., Marliss, E. B., Albisser, A. M., Perl-
man, K., and Caron, D.: Open loop intravenous insulin waveforms
for postprandial exercise in type I diabetes. Diabetes Care 1983;
6:129-34.
10
Bjorntorp, P., Dejounge, K., Sjostrom, L , and Sullivan, L :
Physical training in human obesity II. Effects of plasma insulin in
glucose intolerant obese subjects without marked hyperinsulinemia.
Scand. J. Clin. Lab. Invest. 1973; 32:41-45.
11
Bjorntorp, P., Holm, G., Jacobson, B., Dejounge, K., Lund-
berg, P., Smith, V., and Sullivan, L.: Physical training in human
hyperplastic obesity IV and effects on the hormonal status. Metab-
olism 1977; 26:319-28.
12
Ruderman, N. B., Ganda, O. P., and Johnson, K.: The effect
of physical training on glucose tolerance and plasma lipids in ma-
turity-onset diabetes. Diabetes 1979; 28 (Suppl. 1):89—92.
13
Saltin, B., Lingarde, B. S., Jouston, M., Horlin, R., Nygaard,
E., and Gad, P.: Physical training and glucose tolerance in middle-
age men. Diabetes 1979; 28:30-32.
14
Wallberg-Henriksson, H., Gunnarsson, R., Henriksson, J.,
DeFronzo, R. A., Felig, P., Ostran, J., and Wahren, J.: Increased
peripheral insulin sensitivity and muscle mitochondrial enzymes but
unchanged blood glucose control in type I diabetics after physical
training. Diabetes 1982; 31:1044-50.
15
Caron, D., Poussier, P., Marliss, E. B., and Zinman, B.: The
DIABETES CARE, VOL. 7 NO. 6, NOVEMBER-DECEMBER 1984
effect of postprandial exercise on meal-related glucose intolerance
in insulin-dependent diabetic individuals. Diabetes Care 1982; 5:364-
69.
16
Exercise Testing and Training of Apparently Healthy Individ-
uals: A Handbook for Physicians. New York, American Heart As-
sociation, 1972; 14-15.
17
Astrand, P. O.: Aerobic work capacity in men and women
with special reference to age. Acta Physiol. Scand. 1960; 49:169.
18
Pollock, M. L.: The quantification of endurance training pro-
grammes. Exer. Sport Sci. Rev. 1973; 1.
Young, C. M., Hagan, G. C., Tucker, R. E., and Foster,
W. D.: A comparison of dietary study methods II. Dietary history
vs. seven-day record vs. 24 hour recall. J. Am. Diet. Assoc. 1952;
28:218-21.
Reed, R. B., and Burke, B. S.: Collection and analysis of
dietary intake data. Am. J. Public Health 1954; 44:1015-26.
21
Young, C. M., and Trulson, M. F: Methodology for dietary
studies in epidemiological surveys II—strengths and weaknesses of
existing methods. Am. J. Public Health 1960; 50:803-14.
22
Mayer, J., Marshall, N. B., Vitale, J. J., Christensen, J. D.,
Mosha-Gekhi, M. B., and Stare, F. J.: Exercise food intake and
body weight in normal rats and genetically obese adult mice. Am.
J. Physiol. 1954; 177:544-48.
23
Paffenbarger, R. S., Jr., and Hyde, R. T.: Exercise as protection
against heart attack. N. Engl. J. Med. 1980; 302:1026-27.
24
Peterson, C M . , Jones, R. L , Esterly, J. A., Wantz, G. E.,
and Jackson, R. L.: Changes in basement membrane thickening
and pulse volume concomitant with improved glucose control and
exercise in patients with insulin-dependent diabetes mellitus. Di-
abetes Care 1980; 3:586-89.
25
Kannel, W. B., and Sorlie, P.: Some health benefits of physical
activity. Arch. Intern. Med. 1979; 139:857-61.
26
Froelicher, V. F., and Oberman, A.: Analysis of epidemiologic
studies of physical inactivity as risk factor for coronary artery disease.
In Exercise and Heart Disease. Sonnenblick, E. H., and Lesch, M.,
Eds. New York, Grune & Stratton, Inc., 1977:209-27.
27
Morgan, W. P.: Psychological effects of exercise. Behav. Med.
Update 1982; 4:25-30.
519