INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

IT IS WELL RECOGNIZED THAT the ability of insulin to stimulate glucose uptake is enhanced by a single bout of exercise (21). We have shown that the marked increase in insulin action during exercise is matched by an equally profound increase in insulin-induced hypoglycemic counterregulation in an exercising dog model (28). The effectiveness of this potent counterregulatory response may be important in decreasing the magnitude and frequency of hypoglycemia during a bout of exercise. Skeletal muscle remains more sensitive to insulin for a sustained period after exercise (21). In contrast, the counterregulatory response to insulin-induced hypoglycemia is actually reduced a full day after prolonged exercise (9, 16). This is significant, because this likely contributes to the high risk of hypoglycemia in people with insulin-dependent diabetes (27). Whether hypoglycemic counterregulation during the interval immediately after exercise remains sensitized, as it is during exercise, or blunted, as it is after extended recovery from exercise, is unknown.

These studies test whether sensitization of the counterregulatory response to insulin-induced hypoglycemia is present during the early stages of recovery from prolonged exercise (<4 h after exercise) and the tissue-specific metabolic consequences of any sensitization. Experiments were conducted after 150 min of exercise or an equivalent-duration sedentary period in chronically catheterized, overnight-fasted dogs. Glucose fluxes and the tissue-specific fates of glucose in liver and muscle were assessed using isotopic ([3-³H]glucose, [U-¹⁴C]glucose) and arteriovenous difference (liver, limb) techniques.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Animal maintenance and surgical procedures. Mongrel dogs (n = 32, mean wt 22.9 ± 0.3 kg) of either gender that had been fed a standard diet (Pedigree, Vernon, CA) and Wayne Lab Blox (51% carbohydrate, 31% protein, 11% fat, and 7% fiber based on dry weight; Allied Mills, Chicago, IL) were studied. The dogs were housed in a facility that met American Association for the Accreditation of Laboratory Animal Care guidelines, and the Vanderbilt University School of Medicine Animal Care Committee approved the protocols. At least 16 days before each experiment, a laparotomy was performed with animals under general anesthesia (25 mg/kg pentobarbital sodium). Silastic catheters (0.03 in. ID) were inserted into the vena cava for tracer ([3-³H]glucose, [U-¹⁴C]glucose), indocyanine green (ICG), and insulin infusions. Silastic catheters (0.04 in. ID) were also inserted into the portal vein and left common hepatic vein for blood sampling. Incisions were made in the neck and inguinal regions for the placement of carotid artery and common iliac vein sampling catheters, respectively. The carotid artery was isolated, and a Silastic catheter (0.04 in. ID) was inserted and advanced so that its tip rested in the aortic arch. A Silastic catheter (0.03 in. ID) was introduced into the common iliac vein via a lateral circumflex vein. Exposure of the lateral circumflex vein was achieved with a 2-cm incision in the lower femoral region and dissection of the vein from the subcutaneous tissues. The catheter tip was positioned in the common iliac vein, distal to the anastomosis with the vena cava. The median sacral vein was ligated to prevent dilution from other sites. Verification of catheter placement was made through the abdominal incision site. After insertion, all catheters were filled with saline containing heparin (200 U/ml; Abbott Laboratories, North Chicago, IL), and their free ends were knotted.

Experimental procedures. Protocols are illustrated in Fig. 1. Experiments consisted of a period of moderate-intensity (100 m/min, 12% grade) treadmill exercise or rest (200 to 50 min), a basal sampling period (40 to 0 min), and a euglycemic or hypoglycemic clamp period (0 to 270 min). Primed, constant-rate infusions of [3-³H]glucose (primer of 42 µCi, infusion of 0.30 µCi/min) and [U-¹⁴C]glucose (primer of 24 µCi, infusion of 0.17 µCi/min) were initiated at 120 min and continued throughout the study. In addition, a bolus of sodium [¹⁴C]bicarbonate (0.64 µCi/kg) was given, and an infusion of ICG (0.1 mg · m² · min¹) was initiated at this time. At t = 0 min, an infusion of insulin (1 mU · kg¹ · min¹) was started. Arterial glucose was clamped at euglycemic levels in nine sedentary and six exercising dogs and at moderately hypoglycemic levels in nine sedentary dogs and eight exercising dogs. During euglycemic clamps, [3-³H]glucose was mixed with the cold glucose infusate to minimize changes in [3-³H]glucose specific activity. During hypoglycemic clamps, the [3-³H]glucose was adjusted, with the goal again of maintaining [3-³H]glucose specific activity constant. Arterial samples were drawn at 10-min intervals from t = 40 to 150 min. Portal, hepatic, and common iliac vein samples were drawn at t = 40, 20, 0, 10, 30, 50, 60, 80, 100, 110, 130, and 150 min. External iliac artery blood flow was recorded continuously from the frequency shifts of the pulsed sound signal emitted from the Doppler flow.

View larger version (24K): [in this window] [in a new window]   Fig. 1.   Protocols contained a dye and isotope equilibration period (120 to 40 min), a period of moderate-intensity treadmill exercise or rest (200 to 50 min), and a basal sampling period (40 to 0 min). Insulin was infused from t = 0 to 150 min (1 mU · kg1 · min1), and arterial plasma glucose levels were clamped at euglycemic (~110 mg/dl) or moderately hypoglycemic (~65 mg/dl) levels. There were (n) 9 sedentary animals under euglycemic and hypoglycemic conditions, 6 animals under the exercised euglycemic condition, and 8 under the exercised hypoglycemic condition.

Processing of blood samples. Plasma glucose levels were determined by the glucose oxidase method with a Beckman glucose analyzer (Beckman Instruments, Fullerton, CA). Plasma glucose radioactivity (³H and ¹⁴C) samples were deproteinized with barium hydroxide and zinc sulfate and subsequently placed over Dowex 50W-X8 (Bio-Rad Laboratories, Richmond, CA) and Amberlite (Rohm and Haas, Philadelphia, PA) resins. Resined samples were centrifuged, and the supernatant was evaporated and reconstituted in 1 ml of water and 10 ml of Ecolite+ (ICN, Irvine, CA). Radioactivity was then determined by dual-label liquid scintillation counting (model LS-5000TD, Beckman Instruments. Fullerton, CA). ICG concentration was measured spectrophotometrically (805 nm) in arterial and hepatic vein plasma immediately after experiments. Plasma nonesterified fatty acids (NEFA) were measured spectrophotometrically by use of the kit made by Wako Chemicals (Richmond, VA). Whole blood lactate, alanine, and glycerol concentrations were determined in samples deproteinized with an equal volume of 8% perchloric acid by enzymatic methods (13) on a Monarch 2000 centrifugal analyzer (Instrumentation Laboratories, Lexington, MA). The ¹⁴CO₂ in whole blood was liberated by acidification with hydrochloric acid, trapped on chromatography paper using hyamine hydroxide, and quantified using liquid scintillation counting.

Materials. High-performance liquid chromatography-purified [3-³H]glucose and [U-¹⁴C]glucose (New England Nuclear, Boston, MA) were used as the glucose tracers. ICG was purchased from Hynson, Westcott and Dunning (Baltimore, MD). Glucagon and insulin antisera were obtained from Dr. R. L. Gingerich (Washingtion University School of Medicine, St. Louis, MO), and the standard glucagon and ¹²⁵I-labeled glucagon were obtained from Novo Research Institute (Copenhagen, Denmark). Standard insulin and ¹²⁵I-labeled insulin were obtained from Linco Research (St. Louis, MO). Enzymes used in chemical analyses were obtained from Sigma Chemical (St. Louis, MO) or Boehringer-Mannheim Biochemicals (Mannheim, Germany). Doppler flow probes were obtained from the Instrumentation Development Laboratory, Baylor University School of Medicine.

Calculations. Total glucose appearance (R_a) and disappearance (R_d) were determined using the two-compartment model described by Mari (14) for nonsteady state with [3-³H]glucose as the tracer. Endogenous glucose production (also R_a) was calculated as the total glucose production minus the exogenous glucose infusion rate. The two-compartment approach does not result in the underestimate of R_a values seen under hyperinsulinemic conditions when the modified one-compartment model is used (14). The accuracy of the R_a calculation was further strengthened by using a variable [3-³H]glucose infusion rate during the experimental period to minimize changes in glucose specific activity.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Arterial plasma insulin and glucose and glucose infusion rates. Arterial plasma insulin rose from a basal value of ~6 to ~40 µU/ml during the test period in all protocols (Fig. 2). Arterial glucose levels were clamped at euglycemic (114 ± 7 mg/dl in sedentary dogs and 107 ± 5 mg/dl in exercise dogs) or moderately hypoglycemic (68 ± 3 mg/dl in sedentary dogs and 67 ± 2 mg/dl in exercise dogs) levels (Fig. 2).

View larger version (29K): [in this window] [in a new window]   Fig. 2.   Arterial plasma glucose and insulin concentrations during hyperinsulinemic euglycemic and hypoglycemic clamps. Data are means ± SE; n = 9 animals under sedentary euglycemic and hypoglycemic conditions, 6 under the exercised euglycemic condition, and 8 under the exercised hypoglycemic condition.

1

1

1

1

1

1

View larger version (20K): [in this window] [in a new window]   Fig. 3.   Effect of prior exercise on glucose infusion rates during hyperinsulinemic euglycemic and hypoglycemic clamps. Data are means ± SE; n = 9 sedentary animals under euglycemic and hypoglycemic conditions, 6 under the exercised euglycemic condition, and 8 under the exercised hypoglycemic condition. *Significant difference vs. corresponding hypoglycemic group (P < 0.05); significant difference vs. corresponding sedentary group (P < 0.05).

View larger version (23K): [in this window] [in a new window]   Fig. 4.   Effect of prior exercise on the decrement in glucose infusion rates due to moderate hypoglycemia. #, This variable was calculated as the difference between ~110 and ~65 mg/dl clamps for each condition. Significant difference vs. corresponding sedentary group (P < 0.05).

Arterial plasma glucagon, catecholamine, and cortisol levels. Arterial plasma glucagon was significantly higher in exercised groups compared with sedentary groups before the onset of insulin infusion (Table 1). As anticipated, glucagon was higher during hypoglycemia than during euglycemia whether or not the animals were exercised (P < 0.05). The increase due to hypoglycemia was more in exercised dogs compared with sedentary dogs during the first 50 min (P < 0.05).

Glucose kinetics. Although NHGO tended to be higher before insulin infusion after exercise, differences between groups were not significant. As expected, NHGO was significantly suppressed by insulin in the presence of euglycemia in both sedentary (2.6 ± 0.4 and 0.4 ± 0.2 mg · kg¹ · min¹ in the basal state and at t = 150 min, respectively) and exercised (3.3 ± 0.8 and 0.9 ± 0.5 mg · kg¹ · min¹ in the basal state and at t = 150 min) groups (Fig. 5). NHGO was actually negative (net uptake) in five of the six exercised dogs and two of the nine sedentary dogs during euglycemic clamps. Interestingly, the total suppression of NHGO was ~twofold more in exercised dogs. As expected, hypoglycemia resulted in a relative stimulation of NHGO so that the suppressive effects of insulin on this variable were largely overcome in both sedentary (2.1 ± 0.5 and 1.5 ± 0.3 mg · kg¹ · min¹ in the basal state and at t = 150 min) and exercised (3.2 ± 0.5 and 2.3 ± 0.6 mg · kg¹ · min¹ in the basal state and at t = 150 min) dogs. The stimulation of NHGO due to hypoglycemia (i.e., the difference between NHGO during hypoglycemic and euglycemic clamps) was greater in the postexercise state (Fig. 6).

View larger version (22K): [in this window] [in a new window]   Fig. 5.   Effect of prior exercise on net hepatic glucose output during hyperinsulinemic euglycemic and hypoglycemic clamps. Data are means ± SE; n = 9 sedentary animals under euglycemic and hypoglycemic conditions, 6 under the exercised euglycemic condition, and 8 under the exercised hypoglycemic condition. *Significant difference vs. corresponding euglycemic group (P < 0.05).

View larger version (26K): [in this window] [in a new window]   Fig. 6.   Effect of prior exercise on stimulation of net hepatic glucose output (NHGO) due to a decrement in glucose. This variable was calculated as the difference between ~110 and ~65 mg/dl clamps for each condition. Significant difference vs. corresponding sedentary group (P < 0.05).

1

1

1

1

1

1

1

1

View larger version (22K): [in this window] [in a new window]   Fig. 7.   Effect of prior exercise on endogenous glucose appearance rates during hyperinsulinemic euglycemic and hypoglycemic clamps. Data are means ± SE; n = 9 sedentary animals under euglycemic and hypoglycemic conditions, 6 under the exercised euglycemic condition, and 8 under the exercised hypoglycemic condition. *Significant difference vs. corresponding hypoglycemic group (P < 0.05).

View larger version (26K): [in this window] [in a new window]   Fig. 8.   Effect of prior exercise on the increase in endogenous glucose appearance rates due to a decrement in glucose. #This variable was calculated as the difference between ~110 and ~65 mg/dl clamps for each condition. Data are means ± SE.

1

1

1

1

1

1

1

1

View larger version (22K): [in this window] [in a new window]   Fig. 9.   Effect of prior exercise on glucose disappearance during hyperinsulinemic euglycemic and hypoglycemic clamps. Data are means ± SE; n = 9 sedentary animals under euglycemic and hypoglycemic conditions, 6 under the exercised euglycemic condition, and 8 under the exercised hypoglycemic condition. *Significant difference vs. corresponding euglycemic group (P < 0.05). Significant difference vs. corresponding sedentary group (P < 0.05).

View larger version (27K): [in this window] [in a new window]   Fig. 10.   Effect of prior exercise on the suppression of glucose disappearance rate due to a decrement in glucose. #This variable was calculated as the difference between ~110 and ~65 mg/dl clamps for each condition. Significant difference vs. corresponding sedentary group (P < 0.05).

LGU, LGO, and LGNO. LGU, LGO, and LGNO were not different between groups before the insulin infusion period (Table 2). LGU was increased markedly during euglycemic clamps in both sedentary and exercised dogs (Table 2). This increase was greater in exercised dogs, and it corresponded to a greater rate of LGNO. LGO was also increased during the euglycemic clamps, but the rate was not influenced by prior exercise. LGU, LGO, and LGNO were all reduced during hypoglycemic clamps in relation to euglycemic clamps. Rates of limb glucose uptake and metabolism during hypoglycemic clamps were not influenced by prior exercise.

Arterial concentrations and net balances of lactate. Table 3 shows that arterial lactate levels were not significantly affected by prior exercise or insulin infusion in the presence of a euglycemic or hypoglycemic clamp. It should be noted that, with regard to net tissue balances of lactate, the variability inherent in these measurements is high, and small to moderate differences may go undetected. Net hepatic lactate uptake was not significantly affected by prior exercise. Glycemia was a significant factor in determining net hepatic lactate balance. The liver was a net producer of lactate during euglycemic clamps, whereas during hypoglycemic clamps, the liver was a net consumer of lactate. Differences were significant during the two latter periods of the insulin infusion. Net limb lactate output was not significantly affected by prior exercise. Although there was a tendency for greater net limb output during hypoglycemia compared with euglycemia, differences were insignificant.

Arterial concentrations and net balances of glycerol. Arterial concentrations, net hepatic uptake, and net limb output of glycerol were not significantly affected by prior exercise (Table 4). Concentrations were suppressed significantly by insulin during euglycemic clamps in both sedentary and exercised dogs. Glycerol concentrations exhibited no change during hypoglycemia after exercise but rose by ~40% in dogs that had remained sedentary. Net hepatic glycerol uptake was significantly reduced during euglycemic clamps regardless of whether dogs were sedentary or exercised. Hypoglycemia caused a significant increase in net hepatic glycerol uptake relative to corresponding basal periods and euglycemic clamp periods in both sedentary and exercised dogs. Net limb glycerol output was suppressed during euglycemic clamps relative to corresponding basal periods, but it was unchanged from basal during hypoglycemic clamps.

Arterial plasma NEFA concentrations. NEFA concentrations were similar in all groups before the insulin infusion (Table 5). As expected, insulin suppressed NEFA concentrations under euglycemic conditions. This suppression was equal in exercised dogs and dogs that had remained sedentary. This effect of insulin was reduced under hypoglycemic conditions. NEFA concentrations were significantly higher during hypoglycemic clamps compared with corresponding intervals during euglycemic clamps. Again, differences were not significantly different between exercised dogs and those that had remained sedentary.

Hepatic and limb blood flows. Hepatic and limb (external iliac artery) blood flows were unaffected by prior exercise, hyperinsulinemia, or hypoglycemia (Table 6).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

It is well established that prior exercise causes increases in insulin-stimulated whole body glucose disappearance, muscle glucose uptake, and muscle nonoxidative glucose metabolism (21, 23). The present studies add two important new elements to our knowledge of the effects of prior exercise on insulin-stimulated metabolism. The first is that prior exercise significantly enhances the action of insulin on the liver, as it does on muscle, under euglycemic conditions. Despite the greater suppressive effect of insulin, prior exercise did not influence NHGO during hypoglycemia. The second new finding, therefore, is that in the presence of moderate insulin-induced hypoglycemia, there must be added stimulatory drive to the liver to overcome the enhanced action of insulin at the liver. It is probable that many factors determine these responses, including the actual duration after exercise in which measurements are made. Results obtained from these experiments may be specific for the ~4-h period after prolonged, moderate-intensity exercise.

With regard to the first finding, prior exercise increased the suppression of NHGO by a physiological increment in insulin by ~twofold under euglycemic conditions. This finding was readily detectable despite the inherent variability of the arteriovenous difference method. It is important to recognize that the ability of insulin to suppress R_a was not enhanced. The lack of an effect of prior exercise on the ability of insulin to suppress R_a is consistent with findings in human subjects (18, 23). It is noteworthy that the reduction in R_a during the euglycemic clamp after exercise appeared to wane during the latter stage of the 150-min clamp, whereas the reduction in NHGO was sustained. This suggests that the effect of insulin on hepatic glucose dynamics may be dependent on the time frame after exercise in which measurements are made. Earlier estimates of R_a that were done after exercise in human subjects were difficult to assess because of weaknesses in the tracer method associated with either possible isotopic impurities (17) or errors in the model used to calculate glucose fluxes (14). Using an HPLC-purified product minimized the possibility of isotope contamination in the present study. The error associated with the model used to calculate glucose fluxes was addressed in two ways. First, a two-compartment model that is less dependent on steady-state conditions was used (14) and not the modified one-compartment model used previously (18, 23). Moreover, varying the isotope infusion rate to minimize changes in specific activity reduced model dependency of the glucose flux calculations.

The observation that suppression of NHGO, but not R_a, was greater during hyperinsulinemic, euglycemic clamps after exercise strongly suggests that insulin-induced stimulation of hepatic glucose uptake was increased. This effect of prior exercise was apparent even though insulin levels were actually lower, albeit not significantly so, by ~20%, before and during the insulin infusion. Moreover, this difference was present even though circulating epinephrine levels actually remained higher after exercise (Table 1). Several studies have suggested that the ability of the liver to consume glucose after exercise is improved (7, 8, 12, 15). It is unknown whether this is due to an insulin-dependent or an insulin-independent mechanism. One study demonstrated that prior exercise increased insulin-induced generation of pyruvate dehydrogenase-stimulating activity in the rat in the absence of any effect on insulin binding to the hepatocyte membrane (1). This suggested that exercise enhances hepatic insulin action at a site distal to the hepatic insulin receptor. The present study is the first of which we are aware to show that hepatic insulin action is increased in vivo. The mechanism for the increase in hepatic insulin action remains to be determined. One of the key determinants of the postexercise increase in muscle insulin action is depletion of muscle glycogen stores (3, 21). The exercise protocol used in the present study depletes liver glycogen stores ~70% (26). It is possible that a reduction in liver glycogen stores could be a cause of the enhanced ability of insulin to enhance hepatic glucose uptake under euglycemic conditions.

Although the glucose infusion rate was higher in the postexercise state during the euglycemic clamp, the glucose infusion rates during the hypoglycemic state were virtually identical in exercised and sedentary dogs. This shows that the counterregulatory response during the postexercise period was sufficiently increased to offset the increase in insulin action. As mentioned earlier, after exercise, insulin was ~20% lower during the insulin clamp period (the difference was insignificant) because of reduced preclamp levels. This could counterbalance the effect of the increase in insulin action after exercise. This is unlikely to be an important variable in these studies, however, as increased insulin action was demonstrable during the euglycemic experiments, even though the same insignificant difference in insulin levels was present. Clearly, then, there is a distinct difference between the responses in euglycemic and hypoglycemic clamps with and without prior exercise. The improved counterregulatory response effectively nullified the exercise-induced increase in whole body glucose uptake by actions on both muscle and liver. Table 3 shows that hypoglycemia reduced muscle glucose uptake and nonoxidative metabolism during insulin clamps but that it did it twice as well after exercise. LGO and nonoxidative metabolism were reduced by ~50 µmol/min during hypoglycemic clamps compared with euglycemic clamps in sedentary dogs. After exercise, however, hypoglycemia resulted in an ~100 µmol/min reduction in these variables. Figure 6 showed that hypoglycemia was twice as effective in stimulating NHGO after exercise. This observation, combined with the fact that hypoglycemia was no more effective in stimulating R_a after exercise, indicates that the exercise-stimulated increase in hepatic glucose uptake was negated by the counterregulatory response. Equivalent hypoglycemic responses in R_a in sedentary and exercised dogs, coupled with equivalent increments in hepatic gluconeogenic precursor (lactate and glycerol) uptake, support the premise that prior exercise did not promote the formation of glucose by the liver. The existence of differences in the gluconeogenic response to hypoglycemia due to prior exercise cannot be completely ruled out, however; because not all of the gluconeogenic precursors were measured, the balance of those we did measure was highly variable, and we have no index of intracellular gluconeogenic regulation.

Research in human subjects shows that the response to an insulin-induced hypoglycemia of ~50 mg/dl is blunted a full day after 180 min of moderate exercise (50% maximum oxygen uptake; two bouts of 90 min each) (9). This is consistent with a recent report showing that the counterregulatory response to a stepped hypoglycemic stimulus that achieved a nadir of 45 mg/dl was blunted one day after 120 min of moderate exercise (70% maximum oxygen uptake; two bouts of 60 min each) (16). In this latter study, however, the blunting was somewhat less marked than in the former study. Duration after exercise seems to be an important parameter, because when insulin-induced hypoglycemia was induced beginning 60 min after moderate exercise (~60% maximum oxygen uptake for 60 min), no blunting was observed in healthy human subjects (20). The mechanism for this time-dependent blunting of the counterregulatory response to hypoglycemia has been proposed to be due to the increment in cortisol that occurs with sufficiently intense exercise (4, 5). In the present study, in which the hypoglycemic period was in even closer temporal proximity to exercise, there was a transiently greater arterial glucagon and cortisol response to hypoglycemia. Granted, it is likely that the increase in portal vein glucagon was considerably higher and perhaps more persistent than arterial glucagon during hypoglycemia in exercised dogs compared with sedentary dogs (19, 25). Epinephrine was insignificantly higher during the hypoglycemic period. This trend appears to be related to residual effects of primary exercise rather than to a difference in the response to hypoglycemia. One cannot rule out, however, that acute exercise enhances the sensitivity to adrenergic stimulation (22). The effects of prior exercise on glucose disappearance and muscle glucose metabolism during insulin-induced hypoglycemia were clearly sustained throughout the experimental period. Glucose fluxes and the overall effectiveness of counterregulation were not assessed in the most comparable human study in which hypoglycemia was induced beginning 60 min after exercise (20).

Lipolysis was, as expected, suppressed by insulin during the euglycemic clamps. This was evidenced by the reduction in arterial glycerol and NEFA concentrations and in net limb glycerol output under these conditions. The magnitude of these decrements was unaffected by prior exercise. Hypoglycemia countered the suppressive effects of insulin on lipolysis in both exercised and sedentary dogs. There was no significant difference in arterial glycerol, arterial NEFA, and net limb glycerol output during hypoglycemia in the two groups. Glycerol rose significantly above the preinsulin infusion value, however, only in the sedentary hypoglycemic group. The lipolytic response did not increase above preinsulin infusion levels in the exercised hypoglycemic group. This observation is consistent with an earlier study showing that prior exercise blunted the lipolytic response to hypoglycemia (9). The generally minimal differences in lipolytic indexes between exercised and sedentary dogs are consistent with the fact that the catecholamine response to hypoglycemia was similar in the two groups. Catecholamines are the primary stimulators of the lipolytic response to hypoglycemia (6).

The exercise-induced increase in insulin action persists well after the cessation of the work bout (21). The duration and magnitude of this persistence are influenced by factors relating to nutritional state, fitness, and duration and intensity of the prior work bout (27). Although the response to insulin-induced hypoglycemia after exercise does not appear to be characterized by the same long-term persistence as other aspects of insulin action (9, 16), it still may be possible that the response is influenced by characteristics of the subjects tested and by exercise condition.

In summary, differences between the response to an insulin-induced decrement in arterial glucose and an equivalent duration period of hyperinsulinemic euglycemia were used to assess the effectiveness of defenses against a moderate decrement in arterial glucose shortly after the cessation of exercise. Results show that, when experiments were initiated immediately after prolonged, moderate exercise and values were compared with those of sedentary controls, 1) the glucose infusion rate required to maintain euglycemia, but not hypoglycemia, was higher, indicating that postexercise insulin sensitivity was counterbalanced by a counterregulatory response that was more effective; 2) R_d was greater under euglycemic, but not hypoglycemic, conditions; 3) NHGO, but not R_a, was suppressed more by a hyperinsulinemic euglycemic clamp, suggesting that hepatic glucose uptake was increased; 4) a decrement in glucose completely reversed the more potent effects of insulin on NHGO; and 5) arterial glucagon and cortisol were transiently higher in the presence of a decrement in glucose. In summary, an increase in insulin action that was readily evident under euglycemic conditions shortly after the cessation of exercise was abolished by moderate hypoglycemia. The means by which the glucoregulatory system is able to overcome the increase in insulin action during moderate hypoglycemia is related not to an increase in R_a but to a reduction in insulin-stimulated R_d. The primary site at which this reduction occurs is the liver.