Glucose Counterregulation in Advanced Type 2 Diabetes: Effect of β-Adrenergic Blockade

In type 1 diabetes mellitus (T1DM) with complete loss of β-cell function, plasma insulin levels do not decrease and plasma glucagon levels do not increase in response to falling plasma glucose concentrations; epinephrine responses initially compensate but become frequently impaired, most commonly as a result of episodes of hypoglycemia. Attenuated epinephrine responses in combination with the absent glucagon responses in T1DM cause the syndrome of defective glucose counterregulation, which is associated with an ∼25-fold increased risk for severe hypoglycemia (1).

Similar to T1DM, in insulin-requiring type 2 diabetes mellitus (T2DM), the proportion of plasma insulin that is derived from exogenously administered insulin does not decrease in response to falling plasma glucose levels, as it solely depends on the absorption from subcutaneous injection sites. In addition, endogenous insulin secretion declines less rapidly than normal (2) and plasma glucagon responses are diminished during hypoglycemia, especially as insulin deficiency worsens (3–8). However, epinephrine responses are generally normal or even increased in T2DM (3–9). It has therefore been proposed that in advanced T2DM, similar to T1DM, epinephrine responses can compensate for the impaired glucagon and insulin responses and that attenuated epinephrine responses or actions cause defective glucose counterregulation (7). If true, this would have significant clinical implications, since the counterregulatory actions of epinephrine are predominantly if not exclusively mediated through β-adrenergic receptors, including stimulation of endogenous glucose production (EGP), substrate supply for gluconeogenesis, and activation of lipolysis as well as inhibition of glucose utilization (1). β-Adrenergic blockade, commonly used for the treatment of hypertension, ischemic heart disease, or congestive heart failure, may hence be detrimental for glucose counterregulation in patients with advanced T2DM, as has been well documented for patients with T1DM (10–16).

In the present studies, we used hyperinsulinemic euglycemic-hypoglycemic clamps and tracer techniques to examine counterregulatory glucose kinetics in advanced T2DM and test the hypothesis that these are impaired by β-adrenergic blockade.

We studied 9 subjects with T2DM and 6 nondiabetic control subjects matched for demographic and physical characteristics (Table 1). For inclusion, T2DM subjects had to be insulin-requiring as a preliminary index of advanced β-cell failure. T2DM subjects were excluded for positive GAD antibodies, clinical evidence for autonomic neuropathy, proliferative retinopathy, any episodes of severe hypoglycemia within the last 12 months, or evidence for hypoglycemia unawareness. None of the subjects had liver or kidney disease, coronary artery disease, or anemia. Moreover, none took β-adrenergic antagonists within the last 6 months prior to the study or medications known to affect glucose metabolism (except for oral hypoglycemic agents in the T2DM subjects). All subjects gave informed written consent after the protocol had been approved by the institutional review board of the Phoenix VA Medical Center.

Eligible T2DM subjects participated in three study visits, consisting of an arginine stimulation test for assessment of β-cell function and two euglycemic-hypoglycemic clamp tests for assessment of glucose counterregulation: one with infusion of the nonselective β-adrenergic antagonist propranolol and one with infusion of normal saline. The clamp tests were separated by at least 2 weeks, and the infusions were given in a single-blind randomized crossover fashion. T2DM subjects were withdrawn from oral hypoglycemic agents 4 days prior to each of these visits. Eligible nondiabetic subjects participated in two study visits, consisting of an arginine stimulation test and a euglycemic-hypoglycemic clamp test.

All subjects reported to the Clinical Research Unit at ∼8:00 a.m. after an overnight fast. T2DM subjects received an intravenous infusion of insulin (2.5 mU · kg⁻¹ · min⁻¹) until plasma glucose concentrations decreased to 120–140 mg/dL (6.7–7.8 mmol/L) to avoid priming effects of hyperglycemia on insulin secretion. The insulin infusion was followed by a 60-min insulin washout period, during which plasma glucose levels were monitored. At the end of this period, at a target plasma glucose level of ∼120 mg/dL (6.7 mmol/L), two arterialized venous blood samples were obtained 10 min apart for measurement of baseline plasma glucose and C-peptide concentrations. Subsequently, 5 g arginine hydrochloride (R-Gene 10, 10% arginine hydrochloride; Pharmacia & Upjohn, Kalamazoo, MI) was given intravenously over 30 s. Arterialized venous blood samples were obtained at 2, 3, 4, and 5 min after the injection. Nondiabetic subjects were studied in an analogous fashion except that a 60-min infusion of normal saline was given instead of the infusion of insulin.

Subjects were admitted to the Phoenix VA Medical Center at ∼6:00 p.m. the day before the experiment and received a standard dinner (10 kcal/kg: 50% carbohydrate, 35% fat, and 15% protein) between 6:30 and 7:00 p.m. Subjects fasted thereafter except for water ad lib until the experiments were completed. Starting at ∼10:00 p.m., T2DM subjects received an intravenous infusion of insulin overnight to restore near normoglycemia (target glycemia 90–140 mg/dL [5.0–7.8 mmol/L]) as previously described (17). Blood glucose concentrations were measured at ∼30-min intervals throughout the overnight insulin infusion, of which none were in the hypoglycemic range. Moreover, none of the subjects experienced hypoglycemia symptoms.

At ∼7:00 a.m. the following morning, a primed continuous infusion of [3-³H]glucose (30 µCi, 0.3 µCi/min; PerkinElmer, Shelton, CT) was started for glucose turnover measurements. Thereafter, in the T2DM subjects who underwent the clamp experiment on two separate occasions, a continuous infusion of propranolol (1.4 µg · kg⁻¹ · min⁻¹ after a priming dose of 40 µg/kg) or normal saline was started. At ∼8:00 a.m., all subjects had a dorsal hand vein cannulated in the retrograde manner and the hand placed in a heating pad for sampling arterialized venous blood (18). Starting at ∼10:30 a.m., three baseline blood samples were collected from the dorsal hand vein at 15-min intervals (−30, −15, and 0 min) for measurement of plasma concentrations of glucose, lactate, free fatty acids (FFAs), insulin, C-peptide, glucagon, epinephrine, norepinephrine, growth hormone, and cortisol as well as [3-³H]glucose specific activities. At 0 min, a continuous 4-h infusion of insulin was begun (1.5 mU · kg⁻¹ · min⁻¹). Plasma glucose concentrations were maintained at 5.0 mmol/L during the first 120 min (0–120 min), allowed to drop to ∼2.8 mmol/L over the next 45 min (120–165 min), and then maintained at 2.8 mmol/L until the end of the experiment (240 min). Plasma glucose concentrations were adjusted by a variable intravenous infusion of 20% glucose with plasma glucose concentrations measured at 5-min intervals. The 20% glucose infusion was spiked with [3-³H]glucose to minimize changes in plasma glucose specific activities that might lead to an underestimation of the rate of glucose appearance (19).

During the insulin infusion, arterialized venous blood samples were obtained at 30, 60, 90, 105, 120, 150, 180, 210, 225, and 240 min for measurements as described above. Throughout the study (−30 to 240 min) heart rate and rhythm were continuously monitored (Propaq Encore, Model 202 EL; Protocol Systems, Beaverton, OR).

All analyses were performed as previously described (17,20). Arterialized venous blood samples were collected for plasma insulin, C-peptide, glucagon, cortisol, and growth hormone in EDTA tubes containing a protease inhibitor and for plasma catecholamines in EGTA tubes. Plasma glucose and lactate were immediately determined with a glucose analyzer (YSI 2300 STAT Plus Glucose and Lactate Analyzer; Yellow Springs Instruments). For other determinations, samples were placed immediately in a 4°C ice bath and plasma was subsequently separated by centrifugation at 4°C. For consistency, all samples of a given subject were analyzed together in the same assay. Plasma insulin, C-peptide, glucagon, growth hormone, and cortisol concentrations were determined by standard radioimmunoassay. Plasma epinephrine and norepinephrine concentrations were measured by a radioenzymatic method at the University of Washington as previously described (21). Plasma FFA levels were measured by an enzymatic colorimetric method. For determination of plasma [3-³H]glucose specific activities, plasma was deproteinized with barium and zinc and the protein-free supernatant counted after evaporation to dryness.

Baseline EGP and systemic glucose uptake (SGU) were calculated using the Steele steady-state equation (22). During the insulin infusion, the non–steady-state equation of de Bodo et al. (23) was used to calculate systemic glucose appearance and SGU, and EGP was calculated as the difference between systemic glucose appearance and the glucose infusion rate (GIR) required to maintain the desired glycemic level.

Overall glucose counterregulation was calculated as the difference in the average GIR between the last 30 min of the hypoglycemic clamp (210–240 min) and the last 30 min of the euglycemic clamp (90–120 min). Glucose counterregulation by EGP and SGU were calculated in an analogous manner. Averages of the last 30 min of the euglycemic and the hypoglycemic clamp period were also used for other comparisons.

Unpaired two-tailed Student t tests were used for comparisons of the arginine stimulation test and the euglycemic-hypoglycemic clamp test with saline infusion between the T2DM and nondiabetic control subjects. Paired two-tailed Student t tests were used for comparisons of the clamp test with propranolol infusion and the clamp test with saline infusion in the T2DM subjects. Unless stated otherwise, data are expressed as means ± SEM. P values <0.05 were considered statistically significant. All data are presented as means ± SE.

Plasma glucose concentrations immediately prior to the arginine infusion averaged 7.6 ± 0.3 mmol/L in the T2DM subjects and 5.2 ± 0.2 mmol/L in the control subjects. Despite the T2DM subjects’ slightly higher plasma glucose level, concurrent mean plasma C-peptide concentrations were ∼60% reduced (1.6 ± 0.4 vs. 3.9 ± 0.5 pmol/L, P < 0.01). Mean plasma C-peptide concentrations in response to arginine administration were ∼65% reduced in the T2DM subjects (2.4 ± 0.6 vs. 7.1 ± 0.9 pmol/L, P < 0.01).

Overnight infusion of insulin in the T2DM subjects resulted in modest but not statistically significant increases in baseline plasma insulin concentrations compared with the control subjects and fully normalized baseline plasma glucose concentrations (Fig. 1). T2DM subjects had significantly increased baseline plasma glucagon (75.8 ± 7.9 vs. 50.9 ± 4.7 ng/L, P < 0.05) and epinephrine (349 ± 82 vs. 172 ± 38 pmol/L, P < 0.03) concentrations and tended to have increased plasma norepinephrine (1,684 ± 417 vs. 1,094 ± 111 pmol/L, P = 0.30) and cortisol (297 ± 26 vs. 209 ± 29 nmol/L, P = 0.058) concentrations. Baseline growth hormone levels were similar in both groups (Fig. 2).

During the euglycemic-hypoglycemic clamp, insulin infusion increased plasma insulin to similar extents in the T2DM subjects and the control subjects. Plasma glucose concentrations were virtually identical in both groups (Fig. 1). Plasma glucagon decreased during the euglycemic clamp period and subsequently increased during the hypoglycemic clamp in the T2DM subjects and the control subjects, but this increase was markedly diminished in the former group. In the T2DM subjects, the mean increment of plasma glucagon during hypoglycemia compared with euglycemia was approximately threefold reduced (16.0 ± 4.2 vs. 48.6 ± 6.0 ng/L, P < 0.05), and levels did not exceed baseline in contrast to the control subjects. During the euglycemic clamp, plasma epinephrine (373 ± 35 vs. 169 ± 16 pmol/L in control subjects, P < 0.001) remained significantly increased and plasma cortisol became significantly increased (304 ± 17 vs. 218 ± 22 nmol/L in control subjects, P < 0.02) in the T2DM subjects; plasma norepinephrine continued to be nonsignificantly increased, whereas plasma growth hormone continued to be similar to the control subjects. In response to hypoglycemia, both the absolute and the incremental plasma concentrations of epinephrine, norepinephrine, cortisol, and growth hormone were comparable between both groups (Fig. 2).

At baseline, rates of systemic glucose turnover were similar in T2DM and control subjects. During the insulin infusion, the GIRs required to maintain the target plasma glucose level during the euglycemic clamp and the hypoglycemic clamp were also similar between the T2DM subjects and the control subjects. Accordingly, the decrement in the glucose infusion during hypoglycemia compared with euglycemia was comparable between both groups, indicating normal overall glucose counterregulation in the T2DM subjects. Endogenous glucose release was suppressed to similar extents in the T2DM subjects and the control subjects during the euglycemic clamp. However, in response to hypoglycemia EGP subsequently doubled in the control subjects but remained unchanged in the T2DM subjects, such that the increment in EGP was significantly reduced in the latter group (P < 0.05). In contrast, SGU changed similarly in the euglycemic clamp and the hypoglycemic clamp in both groups, such that the decrement in SGU in response to hypoglycemia was unaltered in the T2DM subjects.

Arterial plasma lactate and FFA concentrations did not differ between the T2DM subjects and the control subjects at baseline or during the entire euglycemic-hypoglycemic clamp (data not shown).

Arterial plasma concentrations of glucose and insulin were virtually identical in the propranolol experiments and the saline control experiments (Fig. 1). Propranolol infusion compared with saline infusion significantly decreased the heart rate at baseline (61.9 ± 2.2 vs. 72.2 ± 2.5 bpm, P < 0.001) and during the euglycemic clamp (61.2 ± 2.2 vs. 73.3 ± 3.0 bpm, P < 0.002) and completely abolished the tachycardic responses to hypoglycemia (58.7 ± 2.1 vs. 79.8 ± 2.6 bpm, P < 0.001). Plasma concentrations of glucagon, epinephrine, norepinephrine, cortisol, and growth hormone were similar throughout both experiments (Fig. 2).

The mean GIR required to maintain the euglycemic clamp was >20% increased in the propranolol experiments compared with the saline experiments (P < 0.01). Subsequently, during the hypoglycemic clamp, GIR decreased similarly in the propranolol experiments and the saline experiments (P > 0.3) (Table 2). In addition, the decrement of the GIRs during hypoglycemia compared with euglycemia did not differ between both experiments, indicating that overall glucose counterregulation was unaltered by β-adrenergic blockade.

In the propranolol experiments, EGP was suppressed during the euglycemic clamp and failed to increase in response to hypoglycemia similar to the saline experiments. In contrast, SGU increased by an ∼20% greater extent during the euglycemic clamp in the propranolol experiments than in the saline experiments (P < 0.01). In both experiments, SGU subsequently decreased to rates similar to those in the hypoglycemic clamp (P > 0.3). Further, the decrement of SGU in response to hypoglycemia was not significantly different (P > 0.2) (Table 2).

At baseline and during the euglycemic clamp, arterial plasma lactate and FFA concentrations were similar in the propranolol experiments and the saline experiments. However, in the propanolol experiments, arterial concentrations of lactate (845 ± 27 vs. 1,116 ± 96 µmol/L, P < 0.02) and FFA (84 ± 8 vs. 118 ± 19 µmol/L, P < 0.01) were significantly decreased during the hypoglycemic clamp.

Patients with T2DM who have advanced β-cell failure are known to have impaired responses of glucagon to hypoglycemia, which normally drive hepatic glucose production (24). Epinephrine responses may hence become critical by promoting hepatic and renal glucose production and limiting SGU through β-adrenergic mechanisms (1,25). Therefore, the aims of the current study were to examine counterregulatory glucose kinetics in such T2DM patients and test the hypothesis that they are impaired by β-adrenergic blockade. Advanced β-cell failure was evidenced by the requirement of exogenous insulin replacement and threefold reduced plasma C-peptide levels in response to arginine stimulation. These markedly insulin-deficient T2DM patients demonstrated approximately threefold reduced counterregulatory glucagon responses but intact counterregulatory responses of epinephrine, norepinephrine, growth hormone, and cortisol. Overall, glucose counterregulation was fully preserved, as judged by the absolute rate of glucose infusion required to maintain the hypoglycemic clamp and by its decrement from the euglycemic clamp to the hypoglycemic clamp. However, the pathways underlying overall glucose counterregulation were altered. In contrast to the control subjects, where EGP nearly doubled during the hypoglycemic clamp compared with the euglycemic clamp, EGP failed to increase in response to hypoglycemia in the T2DM subjects. Therefore, the decrement in the GIR in response to hypoglycemia was exclusively due to decreased systemic glucose disposal, which is to a large extent driven by decreased glucose mass action effects, in the T2DM subjects but due to both increased EGP and decreased systemic glucose disposal in the control subjects. These findings differ from those obtained by Shamoon et al. (5) in patients who had T2DM of shorter duration and markedly less impaired glucagon responses to hypoglycemia. In these individuals, both increased EGP and decreased systemic glucose disposal contributed to overall glucose counterregulation. Together, these observations suggest that the progressive decline in glucagon responses with advancing β-cell failure reduces counterregulatory EGP and hence makes the decrease in systemic glucose disposal in overall glucose counterregulation increasingly important.

Contrary to our hypothesis, infusion of the nonselective β-adrenergic antagonist propranolol did not impair glucose counterregulation in T2DM subjects who have advanced β-cell failure and markedly diminished counterregulatory glucagon responses. Propranolol also did not alter the changes in EGP and systemic glucose disposal in response to hypoglycemia. However, propranolol infusion completely prevented the tachycardic response and significantly reduced plasma FFA levels during hypoglycemia, indicating effective β-adrenergic blockade. Why β-adrenergic blockade had no effect on counterregulatory glucose kinetics is unclear from the present data. One possible explanation for this finding is that glucagon responses were markedly diminished but not absent in our T2DM subjects. These residual glucagon responses might have prevented a further decline in EGP during continued hyperinsulinemia in the hypoglycemic clamp, such that β-adrenergic mechanisms were not critical for glucose counterregulation. This notion is supported by previous studies in patients with T1DM, demonstrating that the magnitude of recovery from insulin-induced hypoglycemia during β-adrenergic blockade correlates with residual glucagon responses (12). Certainly, it is also conceivable that mechanisms other than β-adrenergic stimulation might have compensated for the diminished glucagon responses during hypoglycemia in the T2DM subjects, which may be dissected by the pancreatic-pituitary clamp technique.

In the current study, the T2DM subjects had significantly elevated counterregulatory hormone concentrations at normal plasma glucose concentrations compared with the control subjects: plasma glucagon, norepinephrine, and cortisol were ∼1.5-fold increased, and plasma epinephrine was approximately twofold increased at baseline and during the euglycemic clamp. This is consistent with previous studies by Spyer et al. (9), showing that counterregulatory hormone levels increase when T2DM patients with similar glycemic control are rendered normoglycemic by an intravenous insulin infusion. Our finding that β-adrenergic blockade, which of note has generally been associated with worsened insulin sensitivity (26), increased the GIR required to maintain the hyperinsulinemic-euglycemic clamp in our T2DM subjects extends this observation and provides evidence that the increased β-adrenergic activity that occurs in response to controlling glycemia into the normal range has in fact significant metabolic effects in T2DM patients. In addition, these findings support the notion proposed by Spyer et al. (9) that the increased counterregulatory hormone levels make the achievement of normoglycemia more challenging in clinical practice.

Our findings have both practical and theoretical implications. First, β-adrenergic antagonists are widely used for the treatment of hypertension, ischemic heart disease and congestive heart failure in T2DM patients. The present data indicate that these drugs may increase insulin sensitivity at normoglycemia but may not pose an increased risk of hypoglycemia as a result of impaired glucose counterregulation in T2DM patients, even those with advanced β-cell failure and reduced counterregulatory glucagon responses. Clearly, the latter needs to be confirmed in larger studies reflecting real-life conditions. In addition, the present study provides theoretical insight into the counterregulatory mechanisms involved in advanced T2DM and further support the concept that β-adrenergic mechanisms only become critical if glucagon responses are absent.

In summary, we conclude that in patients with advanced T2DM, overall glucose counterregulation is preserved but the contribution of EGP is diminished. Nonselective β-adrenergic blockade does not impair glucose counterregulation, suggesting that β-blocker therapy may not increase the risk of severe hypoglycemia in these patients.

Acknowledgments. The authors thank Dr. Philip E. Cryer for kindly performing the measurements of plasma catecholamine concentrations. Moreover, the authors thank Jonelle Schmidt and Ginger Brechtel for their excellent nursing help.

Funding. The present work was supported in part by an American Diabetes Association award to C.M.

Duality of Interest. No potential conflicts of interest relevant to this article were reported.

Author Contributions. S.B., E.P., P.E., and A.G. performed the research. C.M. designed and performed the research, analyzed data, and wrote the manuscript. C.M. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Prior Presentation. Part of this work was presented at the 67th Scientific Sessions of the American Diabetes Association, Chicago, IL, 22–26 June 2007.