Diabetologia 9 Springer-Verlag 1982

Transcription

1 Diabetologia (1982) 23 : Diabetologia 9 Springer-Verlag 1982 Absence of Exercise-Induced Hypoglycaeia in Type I (Insulin-Dependent) Diabetic Patients During Maintenance of Noroglycaeia by Short-Ter, Open-Loop Insulin Infusion M. J. Martin t, D. C. Robbins 1' 2, R. Bergenstal ~, B. LaGrange 2 and A. H. Rubenstein ~ 1Endocrinology Section, Departent of Medicine, University of Chicago, Chicago, Illinois and 2 The Metabolic Unit, Departent of Medicine, University of Veront, Burlington, Veront, USA Suary. To assess the risk and possible echaniss ofhypoglycaeia during oderate exercise in Type I (insulin-dependent) diabetic patients receiving constant insulin infusion, five insulin-dependent ale diabetic patients were exercised 18 h after their last eal and 30 h after their last injection of interediate acting insulin. Intravenous insulin was initially delivered via a closed-loop infusion syste prograed to lower ean blood glucose fro 11.3 _+ 1.8 to ol/1 over approxiately 3.5 h. Blood glucose was then aintained at this level for 4 h. At this tie, the dosed-loop infusion was discontinued and replaced by an open-loop syste. The average aount of insulin infused per rain during the 4 h noroglycaeic closed-loop period was calculated and this aount was infused at a constant rate during both a 30 in period of exercise on a bicycle ergoeter (approxiately 65% axiu oxygen uptake) and a 30 in rest period which followed. Five non- diabetic ales served as control subjects. Despite significantly higher free insulin concentrations (p < 0.05) and identical preexercise blood glucose concentrations, blood glucose rose during exercise only in the diabetic group (0.5 _+ 0.2 retool/1; p < 0.01). Changes in the seru concentrations of lactate, glycerol, glucagon, cortisol, non-esterified fatty acids and growth horone were siilar in the two groups and did not account for the increent of blood glucose in the diabetic patients. Betahydroxybutyrate concentrations were, however, higher in the diabetic patients at the onset of exercise (p < 0.01) and decreased significantly ore than the control subjects during exercise. We conclude that exercise under these conditions in diabetic patients is not attended by hypoglycaeia. Key words: Type I diabetes, exercise, open-loop insulin infusion. One ethod of noralizing blood glucose is an openloop insulin infusion pup which delivers insulin at a constant rate with pulses of insulin before eals [1, 2] without iediate regard to the concentration of blood glucose. This therapy ay put patients at risk of hypoglycaeia: those ost at risk would appear to be Type I patients whose exercise and eal plans ay be ost unpredictable. There are several reasons why exercise ight cause a fall in blood glucose in diabetic subjects who receive a constant insulin infusion. Firstly, free insulin concentrations are usually higher in the such patients copared with non-diabetic subjects [3]. Secondly, unlike noral subjects in who insulin release declines during exercise [4], there is no decline in the rate of insulin infusion given via a constant open-loop infusion apparatus. Thus, high insulin levels in cobination with increased glucose consuption could be expected to produce hypoglycaeia. On the other hand, portal vein insulin concentrations produced by these devices ay be low, relative to those in the non-diabetic. Thus, despite increased pc- ripheral concentrations, the aount of insulin reaching the liver ay be insufficient to regulate hepatic glucose production appropriately. Under these circustances, blood glucose ay even rise during oderate exercise. The purpose of this study was to assess the effect of exercise in Type I diabetic patients receiving a constant, open-loop insulin infusion prograed to achieve and aintain noroglycaeia while at rest. Subjects and Methods The study was conducted at the University of Chicago's Clinical Research Centre. Infored consent was obtained fro each subject before the investigation. Five ale Type I diabetic patients (age 30 _ 5 years, duration of diabetes 8 _ 2 years), without clinically detectable coplications, and five age-atched, healthy ale control subjects were studied. Insulin regiens of the diabetic subjects are shown in Table 1. All subjects were accustoed to oderate, erratic, physical exercise but none engaged in regular strenuous, copetitive exercise. The subjects were instructed to eat their noral diet until the evening before the study. Each diabetic took his noral orning insulin on the day before the study, but, thereafter, interediate and long-acting insulins were withheld X/82/0023/0337/$01.20

2 338 M.J. Martin et al.: Exercise in Noroglycaeic Diabetes Table 1. Insulin requireents in the diabetic group Subject Usual orning dose Usual evening dose Insulin dose (withheld before study) (Units. kg -1. d-1) 1 40 U NPH U NPH U NPH; 5 U CZI a 5 U NPH; 5 U CZI U NPH; 6 U CZI 12 U NPH; 6 U CZI U NPH 0.74 a crystalline zinc insulin Exercise To deterine the ergoeter tension that would give a pulse rate of approxiately 150 beats/rain when pedalling at kin/h, each subject exercised before the study on a stationary bicycle ergoeter for 7 in. This rate of exercise corresponds to approxiately 65% of axiu oxygen uptake [5]. This tension and pedalling rate were kept constant throughout the 30-in exercise period. During rest periods, subjects were in a sitting position. Glucose Controlled Insulin Infusion Syste The artificial B cell [Biostator, glucose-controlled insulin infusion syste], Life Science Instruents, Miles Laboratories, Elkhart, Indiana, USA], and the regulating algoriths have been described fully elsewhere [6, 7]. Briefly, blood is continuously withdrawn at a rate of 2 l/h through a double luen cannula placed in an antecubital vein. On the basis of continuous whole blood glucose analyses, the achine infuses insulin and/or glucose to aintain noroglycaeia. In this experient only insulin was infused. Experiental Protocol After an overnight fast of at least 12 h and 24 h after the last injection of interediate acting insulin, three indwelling cannulae were inserted into ar veins of each diabetic subject. After 15 in rest, baseline blood saples were obtained. The glucose-controlled insulin infusion was started and blood glucose concentrations were decreased to approxiately retool/1. The average tie required to noralize the blood glucose was 3.5 h. Blood glucose concentrations were aintained at this level for at least I h when a second set of blood saples was obtained. Blood glucose was then aintained at this level for 4 h by infusion of insulin according to the glucose-controlled insulin infusion syste algorith. The average rate of the insulin infusion needed during these 4 h was calculated (68 +_ 25 uu 9 kg -1 9 in -1) and this aount was then given throughout the 30-rain exercise and post-exercise rest period as an open-loop, constant-rate infusion. Non-diabetic control subjects were studied after an overnight fast during a siilar protocol. They were not attached to the glucose-controlled insulin infusion syste but had one indwelling cannula in a forear vein to facilitate blood collection. A volue of noral saline was infused equal to that adinistered to the diabetic group. Analytical Methods Glucose concentrations were deterined by kinetic assay using hexokinase [8]. Free insulin and C-peptide [9] were deterined by radioiunoassay. Glucagon was assayed using 30K antiseru [10]. A coloroetric ethod was used to easure non-esterified fatty acid concentrations [11]. Enzyatic ethods were used to deterine lactate,/3-hydroxybutyrate and glycerol concentrations [12]. Growth horone [13] and cortisol [14] were deterined using radioiunoassay and copetitive protein binding assay, respectively. The sensitivities and intra-assay coefficients of variation for several of the assays were, respectively: insulin 1.0 U/1, 5.7%; C-peptide 0.02 pol/l, 7.5%; glucagon 5 pg/l, 8.8%; fl-hydroxybutyrate 20 ~ol/1, 4.4%; glycerol 20 ~ol/1, 4.5%; growth horone 0.07 ng/ l, 4.0%; cortisol 0.1 nol/1, 2.0%. Statistical Methods Results are expressed as ean +_ SEM. Significance of the difference between two ean values was calculated using the Student's t-test. The ean change between any two tie points was calculated by averaging the changes for each subject (paired t-test). Results To achieve euglycaeia in the diabetic subjects, an average of _+ 932 U of insulin was required. Once euglycaeia was achieved, 1065 _+ 350 U of insulin ( lxu 9 kg- ~ 9 rain- 1) were required to aintain the plasa glucose concentrations between 4.4 and 5.0 ol/1. There were no significant differences in heart rates between the noral and diabetic subjects at any tie point (Table 2). The ost rapid pulse for both diabetic (154 +_ 8 beats/in) and control subjects ( beats/ rain) occurred at the end of the exercise period. Blood glucose values during the study are shown in Figure 1. At the onset of the experient the blood glucose concentrations of the diabetic patients ( ol/1) were significantly higher than the noral subjects (5.1 _+ 0.1; p < 0.02). Thereafter, there were no significant differences in blood glucose levels between the two groups before exercise. However, the groups differed in their response to exercise. The ean rise in blood glucose for the diabetic patients (0.6 _+ 0.2 ol/1) reached statistical significance (p < 0.01), while that of the control subjects (0.1 +_ 0.1 ol/1) did not. These changes were also significantly different when the two groups were copared with each other (p < 0.02). During the 30 rain of rest following exercise, plasa glucose continued to rise in the diabetic group, but only slightly (0.2 _+ 0.1 ol/1). The control group had a gradual but significant decline in the concentration ofiunoreactive free insulin during the pre-exercise period (p < 0.05) (Table 2). A further decline was observed during exercise. Free iunoreactive insulin concentrations in the diabetic group were several-fold higher at all tie points co-

3 M. J. Martin et al.: Exercise in Noroglycaeic Diabetes 339 pared with the control group and were constant through- 14z ~- out the experiental period (p < 0.05). The non-esterified fatty acid (NEFA) concentrations (Fig. 2) were higher in the diabetic subjects at all tie points, but the difference between groups was significant only before noralization of blood glucose. The NEFA concentrations in both groups increased during the period before exercise. During exercise, NEFA concentrations declined during the first 10 in, then rose in both groups. The late rise in NEFA concentrations seen during the last 20 rain of exercise tended to be greater in the 6.0 diabetic patients ( ol/l ;p < 0.02) than the control subjects ( ol/1; p < 0.01), but the difference between the groups was not statistically sig- O 4.0 nificant. The decline in NEFA concentration during the first 10 in of exercise was not significant, fl-hydroxybutyrate concentrations (Fig.2) were arkedly higher in the diabetic patients at all ties, There was a gradual rise in the concentration before exercise in both the diabetic ( ol/1; p < 0.01) and control (0.10 _ 0.05 ol/1) groups. The agnitude of the increase in the diabetic patients was greater than in the control group (p < 0.02). During exercise, fl-hydroxybutyrate concentrations initially decreased in both groups. The ean fall was uch larger in the diabetic ( ol/1; p < 0.001) than the control group (0.07 _ ol/1; p < 0.05). Furtherore, these changes were significantly different fro one another(p < 0.02). There were no significant differences in the glycerol concentrations between the two groups (Fig.2). The ean rise during exercise tended to be higher in the diabetic ( ol/1; p < 0.001) than the control 120 < 'o-oi\ "5 a.o o z ~//t// 240,,,, , Tie ( rain ) [CLOSED I CLOSED ICOI~TAIqI'I LOOP LOOP INSULIN INSULIN INFUSION INFUSION INFUSION BY TO TO OPEN ACHIEVE MAINTAIN LOOP NORMO- NORMO- PROTOCOL GLYCAEMIA GLYCAEMIA Fig. 1. Blood glucose concentrations in the diabetic patients (0--0) and control subjects (O--- O) are shown. The 30-rain period of exercise beginning at 240 in and terinating at 270 in is arked by the horizontal bar. A significant difference between the groups before exercise occurred only in the first easureent (t7 < 0.02). The ean rise in blood glucose in the diabetic patients during exercise was significant (p < 0.01) while no change was disceable in the control group. These exercise-associated changes were different when the two groups were copared with each other (p < 0.02) Table 2. Heart rate and concentration of various substances in diabetic patients and control subjects Begin closed-loop infusion or 0.85% saline Rest Maintain noroglycaeia (closed-loop infusion) Tie (rain) Exercise (Open-loop infusion) Tie (rain) Rest Heart rate Diabetic 66 _+ (beats/rain) Control 60 _+ Lactate Diabetic (ol/l) Control Growth Diabetic 4 _+ horone Control 1 (ng/l) Cortisol Diabetic 221 _ _ (nol/l) Control 295 _ _ _+61 Glucagon Diabetic 92 _ _ _ (pol/l) Control 123 _ _ _ _+ 9 Insulin Diabetic 22.5 _+ (U/l) control Glucose Diabetic (ol/l) Control 5.1 _+ Results expressed as ean + SEM; a Indicatesp < 0.05 between groups _+ I _ _ _ _ _ ~ 19.5 _ 3.8 a 18.7 _+ 7.3 a 19.8 _+ 5.6 ~ 20.6 _+ 4.7 ~ 17.4 _+ 4.0 a 19.3 _+ 4.1 a 19.9 _+ 5.7 a _ _ _ _ a 4.8 _ _ _ _ _

4 340 M.J. Martin et al.: Exercise in Noroglycaeic Diabetes A 1.5 O E 1.o E,< LL. 0.5 LU z "~ 0.6 ~ E E v 0.6 ~ 0.4 X ~ 0.2 "10 e- l ~, o I // I //,i,, ~ // T // :L.U _... lr.~'~" " ~"~,~/~ I //~o // n,,,, -T Tie (in) I CLOSED I CLOSED JCONSTANT I LOOP LOOP INSULIN INSULIN INSULIN INFUSION INFUSION INFUSION BY TO TO OPEN ACHIEVE MAINTAIN LOOP NORMO- NORMO- PROTOCOL GLYCAEMIA GLYCAEMIA Fig.2. Seru concentrations ofnon-esterified free fatty acids (NEFA), glycerol and/3-hydroxybutyrate in the diabetic patients (9149 ) and control subjects (9149 are shown. * p < 0.05; ** p < 0.01 significant differences between groups groups ( ol/1; p < 0.05), but the difference between the groups was not significant. There were no detectable differences between the two groups at any tie in concentrations of lactate (Table 2). There was a significant ean rise during exercise in both the diabetic (3.3 +_ 0.6 ol/1; p < 0.01) and control groups ( ol/1; p < 0.01). These changes, however, were not significantly different fro one another. After the 30-in post-exercise rest period, lactate concentrations in both groups returned to preexercise values. In one of the five diabetic patients a circulating antiglucagon antibody prevented easureents of plasa glucagon. In the others pre-exercise glucagon concentrations tended to be lower than those of the control subjects (Table 2), but the difference between the groups was not significant. In both groups, plasa glucagon concentrations rose during exercise, but the increent was not significant. C-peptide concentrations in the diabetic patients (data not shown) were at or below the detection liits of the assay (0.02 pol/1) throughout the protocol. The pre-exercise concentrations of growth horone tended to be higher in the diabetic subjects, although the differences were not significant (Table 2). There was a significant rise in the growth horone concentrations during exercise in both the diabetic ( ng/l; p < 0.05) and the control groups ( ng/l; p < 0.01). Seru cortisol concentrations at each tie point were not significantly different between the two groups (Table 2). There was a significant increase during exercise in both the diabetic (193 +_ 63 nol/l; p < 0.05) and control groups ( nol/1;p < 0.02). Discussion Although the blood glucose in noral subjects can increase by 1.1 to 1.7 ol/1 with vigorous exercise [4] and decrease 0.6 to 2.2 ol/1 with exercise continuing beyond 90 in [15, 16], there is little change during shortter, ild or oderate exercise [4, 17]. This is due to the equalization of glucose production and utilization. In diabetic subjects and experiental anials, the concentration of circulating insulin present during exercise is an iportant factor in deterining whether blood glucose will rise, fall or reain constant during exercise. A sall, but critical concentration of insulin is essential for stiulation of glucose utilization during exercise [18, 19]. When diabetic dogs were subjected to exercise 27 h after their last insulin injection, blood glucose rose sharply due to increased glucose production without a concoitant increase in glucose utilization by exercising uscle [20]. Siilar elevations in plasa glucose concentrations have also been reported in insulin deficient diabetic patients during exercise [21-23]. In contrast to the insulin-deficient odel of exercise, over-insulinized diabetic anials exhibit a fall in plasa glucose during exercise. When depancreatectoized dogs are given a subcutaneous injection of insulin before exercise, plasa insulin levels are elevated, and there is a significant fall in plasa glucose concentrations due to a isatch of glucose uptake (which was siilar to noral dogs during exercise) and hepatic glucose production (which was decreased fro that found in noral dogs) [24-26]. Zinan et al. [27] deonstrated this sae effect in diabetic subjects receiving subcutaneous insulin injections before exercise. Thus, appropriate insulin concentrations during exercise allow optial glucose utilization by peripheral tissues and appropriate regula-

5 M. J. Martin et al.: Exercise in Noroglycaeic Diabetes tion of hepatic glucose production. There is then equilization of glucose production and utilization during exercise, and blood glucose concentrations are relatively stable [28, 29]. In support of this concept, Zinan et al. [30] have shown that exercising diabetic patients have stable blood glucose concentrations during constant insulin infusion which produces blood glucose concentrations of 6.7 ol/1. Under these circustances, glucose production and disappearance rose synchronously and were siilar to the responses observed in noral subjects. In light of these observations in the exercising diabetic patients, the concentrations of free iunoreactive insulin achieved in this study are critical in interpreting the glycaeic response we have described. Obviously, adequate aounts of insulin were infused to produce stable euglycaeia in each diabetic subject while at rest. Furtherore, the concentrations of free iunoreactive insulin in the diabetic subjects were siilar to those reported by others under siilar circustances [3]. Yet, despite apparently adequate insulinization, the glycaeic response during exercise reflects a ild ibalance between glucose production and consuption. Our data do not perit us to deterine whether glucose production was inadequately suppressed, peripheral glucose consuption relatively diinished, or whether both echaniss contributed to the glycaeic response. The rise in plasa glucose in the diabetic subjects during exercise ay have been due to greater aounts or greater effects of circulating stress horones. Those we easured (glucagon, cortisol, growth horone), however, were not different in the control and diabetic groups, at the beginning or end of exercise. Although peripheral concentrations do not necessarily reflect portal concentrations or tissue sensitivity, these data offer no support for the contention that these horones accounted for the observed rise in plasa glucose. A possible role for catecholaines cannot be excluded. Gluconeogenic precursors or other non-carbohydrate fuels ay have also contributed to the exercise-induced glycaeia. In particular, our data suggest relatively greater consuption of fl-hydroxybutyrate by the diabetic group. The pattern of response in both groups was siilar and consisted of a decline in fl-hydroxybutyrate concentration during exercise and an increase after exercise. The pattern is consistent with the observations of other investigators [29, 31]. The diabetic patients, however, had arkedly higher levels offl-hydroxybutyrate throughout the experient and a relatively greater decrease during exercise suggesting increased utilization of ketone bodies as an energy substrate [32, 33]. Higher seru NEFA concentrations and coparatively lower portal to systeic insulin ratios (as a consequence of intravenous, rather than intraportal infusion of insulin) ay explain this observation. An increased utilization of fl-hydroxybutyrate in the diabetic group ay have led to diinished glucose utilization and, at least in part, contributed to the exercise-induced rise in blood glucose. The higher NEFA concentrations found in the diabetic group have been reported by other investigators [21, 29]. Fasting probably caused the NEFA levels to rise in both groups before exercise while the effect of the insulin infusion tended to narrow the difference between the groups. The concentrations of glycerol before exercise and the response of this substrate during exercise was siilar in both groups and consistent with results previously reported for noral subjects [17, 31]. Other studies have deonstrated an abnorally high lactate concentration during exercise in diabetic subjects [21, 29, 31]. During exercise with constant insulin infusion, Murray et al. [31] reported lactate levels in diabetic patients twice as high as in healthy controls. There were, however, no differences in the lactate concentrations recorded during exercise in the two groups in the present study. This ay have been due to the achieveent of noroglycaeia. Open-loop insulin infusion pups can aintain near-noral blood glucose levels in Type I diabetic patients at rest or during oderate activity [1, 2]. In this study, we have extended the results of Zinan et al. [30] to show that short-ter, continuous, intravenous insulin adinistration does not produce hypoglycaeia in fasting patients during or iediately after exercise, even when noroglycaeia is aintained. Further study is needed before the results of this experience can be extended to diabetic subjects undergoing chronic insulin infusion therapy and stresses other than exercise. Adaptation of uscle, fat or liver to uch ore prolonged aintenance of euglycaeia ay result in different responses. Nevertheless, the assuption that hypoglycaeia will be an invariable occurrence when noroglycaeic diabetic patients engage in vigorous exercise or are subjected to stress is certainly not warranted. References 1. Taborlane WV, Sherwin RS, Genel M, Felig P (1979) Reduction to noral of plasa glucose in juvenile diabetes by subcutaneous adinistration of insulin with a portable infusion pup. N Engl J Med 300: Taborlane WV, Sherwin RS, Genel M, Felig P (1979) Restoration of noral lipid and aino acid etabolis in diabetic patients treated with a portable insulin-infusion pup. Lancet 1 : Rizza RA, Gerich JE, Hayond MW, Westland RE, Hall LD, Cleens AH, Service FJ (1980) Control of blood sugar in insulin dependent diabetics: Coparison of an artifical endocrine pancreas continuous subcutaneous insulin infusion, and intensified conventional therapy. N Engl J Med 303: Felig P, Wahren J (1975) Fuel hoeostasis in exercise. N Engl J Med 293 : Astrand PO, Rhying I (1954) A noogra for calculation of aerobic capacity (physical fitness) fro pulse rate during subaxial work. J Appl Physiol 7: Santiago JV, Cleens AH, Clarke WL, Kipnis DM (1979) Closedloop and open-loop devices for blood glucose control in noral and diabetic subjects. Diabetes 28 : Cleens AH, Chang PH, Myers RW (1977) The developent of Biostator, a glucose controlled insulin infusion syste (GCIIS). Hor Metab Res Suppl 7:

6 342 M.J. Martin et al.: Exercise in Noroglycaeic Diabetes 8. Slein MW (1963) D-Glucose: Deterination with hexokinase and glucose-6-phosphate dehydrogenase. In: Bergeyer HU (ed) Methods of enzyatic analysis, 1 st edn, Acadeic Press, New York, pp Starr JI, Horwitz DL, Rubenstein AH, Mako ME (1979) Insulin, proinsulin, and C-peptide. In: Jaffe BM and Behran HR (eds) Methods of horone radioiunoassay, 2nd edn, Acadeic Press, New York, pp Unger RH, Aguilar-Parada E, Muller WA, Eisentraut AM (1970) Studies of pancreatic alpha cell function in noral and diabetic subjects. J Clin Invest 49: Mikac-Devic D, Stankovic H, Boskovic K (1973) A ethod for deterination of free fatty acids in seru. Clin Chi Acta 45: Varley H (1967) Practical clinical biocheistry. London and Interscience Books, New York pp 618~ Odell WD, Rayford PL, Ross GT (1967) Siplified, partially autoated ethod for radioiunoassay of huan thyroid-stiulating, growth, leuteinizing and follicle stiulating horone. J Lab Clin Med 70: Murphy BEP (1967) Soe studies of the protein binding of steroids and their application to the routine icro and ultraicro easureent of various steroids in body fluids by copetitive proteinbinding radioassay. J Clin Endocrinol Metab 27: Ahlborg G, Felig P, Hagenfeldt L, Hendler R, Wahren J (1974) Substrate turnover during prolonged exercise in an. J Clin Invest 53 : Young DR, Pelligra R, Shapira J, Adachi RR, Skrettingland K (1967) Glucose oxidation and replaceent during prolonged exercise in an. J Appl Physio123: Wahren J, Felig P, Ahlborg G, Jorfeldt L (1971) Glucose etabolis during leg exercise in an. J Clin Invest 50: Berger M, Hagg SA, Ruderan NB (1975) Glucose etabolis in perfused skeletal uscle. Interaction of insulin and exercise on glucose uptake. Bioche J 146: Standl E, Janka HU, Dexel T, Kolb HJ (1976) Muscle etabolis during rest and exercise: Influence on the oxygen transport syste of blood in noral and diabetic subjects. Diabetes 25 (Suppl.2): Vranic M, Wrenshall GA (1969) Exercise, insulin and glucose turnover in dogs. Endocrinology 85: Berger M, Berchtold P, Cuppers H J, Drost H, Kley HK, Muller WA, Wiegelann W, Zierann-Telschow H, Giles FA, Kruskeper HL, Zieran H (1977) Metabolic and horonal effects of uscular exercise in juvenile type diabetics. Diabetologia 13: Dorchy H, Niset G, Oos H, Poortans J, Baran D, Loeb H (1977) Study of the coefficient of glucose assiilation during uscular exercise in diabetic adolescents deprived of insulin. Diabete Metab (Paris) 3: Richardson R (1934) Factors deterining the effect of exercise on blood sugar in the diabetic. J Clin Invest 13 : 699 (Abstract) 24. Kawaori R, Vranic M (1977) Mechanis of exercise-induced hypoglyceia in depancreatized dogs aintained on long-acting insulin. J Clin Invest 59: Vranic M, Kawaori R, Wrenshall GA (1974) Mechanis of exercise-induced hypoglyceia in depancreatized insulin treated dogs. Diabetes 23 (Suppl. 1): 353 (Abstract) 26. Vranic M, Kawaori R, Wrenshall GA (1975) The role of insulin and glucagon in regulating glucose turnover in dogs during exercise. In: Horonal Responses in Exercise Syposiu. Med Sci Sports 7: Zinan B, Murray FT, Vranic M, Albisser AM, Liebel BS, McClean PA, Marliss EB (1977) Glucoregulation during oderate exercise in insulin treated diabetes. J Clin Endocrinol Metab 45: 6414i Lyngste J, Clausen J, Dagard S, Levin-Nielsen S, Rehfeld J, Schaffalitzky de Muckadell O, Sestoft L, Trap-Jensen J (1973) Metabolic fuels of uscle and liver in exercising noral subjects and diabetes. Excerpta Med Int Congr Set 280:97 (Abstract) 29. Wahren J, Hagenfeldt L, Felig P (1975) Splanchnic and leg exchange of glucose aino acids and free fatty acids during exercise in diabetes ellitus. J Clin Invest 55: Zinan B, Vranic M, Atbisser AM, Leibel BS, Marliss EB (1979) The role of insulin in the etabolic response to exercise in diabetic an. Diabetes 28 (Suppl. 1): Murray FT, Zinan B, McClean PA, Denoga A, Albisser AM, Leibel BS, Nakhooda AF, Stokes EF, Marliss EB (1977) The etabolic response to oderate exercise in diabetic an receiving intravenous and subcutaneous insulin. J Clin Endocrinol Metab 44: Rennie M J, Park DM, Sulaian WR (1976) Uptake and release of horones and etabolites by tissues of exercising leg in an. A J Physio1231 : Hagenfeldt L, Wahren J (1968) Huan forear uscle etabolis during exercise. III. Uptake release and oxidation of B-hydroxybutyrate and observations on the B-hydroxybutyrate/acetoacetate ratio. Scand J Clin Invest 21 : Received: 1 October 1981 and in revised for: 26 May 1982 David C. Robbins, M. D. Metabolic Unit, Departent of Medicine University of Veront Burlington, VT , USA