Metabolic and Cardiovascular Effects of Carvedilol and Atenolol in Non-Insulin-Dependent Diabetes Mellitus and Hypertension

Methods

Participants

Our research protocol was approved by our institutional review board, and informed consent was obtained from patients before participation. Men and women who had non–insulin-dependent diabetes mellitus and had a supine diastolic blood pressure of 90 to 105 mm Hg on at least two occasions at the end of a 4- to 6-week placebo run-in period were eligible to participate. All patients were referred from the outpatient department of our institution and were consecutively chosen. A total of 45 patients met the inclusion criteria and were randomly assigned to treatment. All but 3 patients completed the study.

Study Design

Our study had a randomized, double-blind design for parallel study groups. Patients who had blood pressure greater than 160/90 mm Hg or who were taking antihypertensive drugs entered a 4- to 6-week run-in period, during which placebo was given to replace the previous antihypertensive drug, if any. Routine hematologic and blood chemistry analyses (hematologic indices, serum sodium and potassium concentrations, liver enzyme levels, urea concentrations, and creatinine concentrations) were done during the initial screening and after the treatment period. Patients were randomly assigned to receive either carvedilol (25 mg once daily) or atenolol (50 mg once daily) in the morning for 24 weeks. After 4 weeks, patients whose diastolic blood pressure while seated was more than 90 mm Hg and had not decreased by at least 10 mm Hg had their dose of study medication doubled for the remaining 20 weeks of the study.

A person who was not involved in trial management randomly assigned the patients using random numbers derived from published tables. The list of randomization numbers was used to label the drug boxes, which were given to the participants sequentially. Both patients and caregivers were blinded to treatment, and randomization codes were not broken until all laboratory measurements had been done. Cardiovascular and metabolic variables were checked at the end of the placebo period and at the end of the active treatment period. All participants were instructed to follow a weight-maintaining diet (50% carbohydrates, 30% lipids, 20% protein) for 3 days before the experiments were done. Side effects, concomitant diseases, and blood pressure were assessed by interview and physical examination every fourth week during treatment.

Clinical and Laboratory Measurements

Patients were asked to refrain from smoking and to fast overnight before each metabolic assessment was done. The euglycemic clamp technique was used to estimate insulin sensitivity in vivo by infusing insulin (1 mU/kg of body weight per minute) and glucose to keep plasma glucose levels at the baseline concentration [8]. At the unchanged plasma glucose concentrations, the amount of glucose required to maintain euglycemia equals whole-body glucose disposal and is expressed in µmol/kg per minute (M). The insulin sensitivity index (M/insulin level during the clamp procedure) measures how effectively plasma insulin induces glucose uptake in insulin-sensitive tissues, such as muscle and fat. Substrate oxidation was estimated by indirect calorimetry [9]. On the day after calorimetry, the patients had an oral glucose tolerance test (75 g of glucose). On the third day, they had an insulin tolerance test (0.15 U/kg).

Blood pressure was measured with appropriate cuff size three times after patients rested for 5 minutes in the supine position. Plasma glucose, insulin, glucagon, and epinephrine levels were measured as described elsewhere [10]; hemoglobin A1c (HbA1c) was measured by column chromatography (Bio-Rad, Milan, Italy); and cholesterol, triglyceride, and high-density lipoprotein cholesterol levels were determined enzymatically [9]. Left ventricular mass normalized by surface area was measured by echocardiography [9]. Serum levels of lipid peroxides were measured as reaction products of malondialdehyde with thiobarbituric acid (thiobarbituric-acid-reactive substances) according to the method of Waravdekar and Sadlaw [11], with slight modifications. Normal lipid peroxide values for our laboratory are 0.34 to 0.86 µmol/L.

Statistical Analysis

All values in the tables are presented as the mean ±SD unless otherwise noted; 95% CIs are provided where appropriate. The areas under the glucose and insulin curves were calculated by trapezoidal rule [12]. Change was calculated as the value obtained at the end of intervention minus the value obtained at the beginning of intervention. A preliminary analysis of variance was used to assess the significance within and between groups. One-sample t-tests were used to compare values obtained before and after carvedilol or atenolol therapy, and two-sample t-tests were used for between-group comparisons.

Results

Three patients (two in the atenolol group and one in the carvedilol group) were unavailable for follow-up; they refused to complete the study and did not specify a reason. These patients dropped out early (between weeks 4 and 8) after randomization; analysis of the study results did not differ when the analysis was done according to actual treatment or according to intention to treat (we used the latter method). Compliance, determined by tablet count, was 94.5% in the carvedilol group and 95% in the atenolol group. The baseline characteristics of the 45 patients who completed the study are shown in Table 1. The two groups were similar at baseline. Body mass index did not change in either group after treatment. Approximately one third of patients in each group (32% in the carvedilol group and 35% in the atenolol group) required upward dose titration at week 4 because of inadequate response. At the end of treatment, 91% of patients receiving carvedilol and 85% of those receiving atenolol had a diastolic blood pressure while seated of less than 90 mm Hg or had their diastolic blood pressure decreased by more than 10 mm Hg (P > 0.2 for comparison). Average systolic and diastolic blood pressure and left ventricular mass decreased in both groups, but the differences between the groups were small (P > 0.2) (Table 2). The decrease in heart rate was greater in patients receiving atenolol than in those receiving carvedilol (P < 0.005). The decrease in mean triglyceride level was 0.56 mmol/L greater (P < 0.001) and the increase in high-density lipoprotein cholesterol level was 0.13 mmol/L greater (P < 0.001) with carvedilol than with atenolol.

Mean fasting plasma glucose and insulin levels decreased during carvedilol treatment and increased during atenolol treatment (Table 2). The HbA1c level decreased by 1.4% in the carvedilol group and increased by 4% in the atenolol group (P < 0.001 for the difference). Mean total glucose disposal and insulin sensitivity index increased during carvedilol treatment and decreased during atenolol treatment (P ≤ 0.01 for the difference). Serum levels of thiobarbituric-acid-reactive substances decreased by 0.25 µmol/L in the carvedilol group and did not change in the atenolol group (P < 0.001 for the difference).

The decreases in plasma glucose and insulin responses to the oral glucose load were 61 mmol/L × 180 minutes greater (CI, −101 to −21 mmol/L × 180 minutes; P = 0.035) and 6.2 nmol/L × 180 minutes greater (CI, −9.8 to −2.6 nmol/L × 180 minutes; P = 0.03), respectively, with carvedilol than with atenolol (data not shown). The plasma glucose level nadir occurred 60 minutes after the insulin bolus was administered and was not affected by either drug (P = 0.09) (data not shown). Glucagon and epinephrine responses to hypoglycemia were similar before and after treatment with both drugs (P > 0.08) (data not shown).

Discussion

Our results show that both carvedilol and atenolol effectively decrease blood pressure and ventricular mass in patients with diabetes and hypertension. They also show that the drug doses administered in this study are equivalent with regard to their ability to decrease blood pressure in these patients, who are particularly at risk for cardiovascular disease. The similarities between the two drugs end with their cardiovascular effects, however; their metabolic effects are different and, to a large extent, divergent. Fasting plasma glucose and insulin levels decreased during carvedilol treatment and increased during atenolol treatment. This indicates that carvedilol improves insulin sensitivity, whereas atenolol weakens it. This finding also fits with the discovery that glucose disposal increased by 20% during carvedilol treatment but decreased by 16% during atenolol treatment.

The divergent effects of carvedilol and atenolol on insulin action may be due to the different effects of these drugs on blood flow to the large muscles of the legs. Both atenolol and metoprolol reduce peripheral blood flow and the availability of glucose to the skeletal muscle [13]. Carvedilol produces peripheral vasodilation caused by α₁-adrenergic blockade, which may facilitate glucose uptake by muscle cells. Other antihypertensive drugs that are known to increase peripheral blood flow (such as captopril, prazosin, and doxazosin) increase insulin sensitivity [14].

The divergent properties of carvedilol and atenolol on peripheral vascular resistances may also explain, at least in part, the significant changes of lipid metabolism associated with their use. Atenolol caused the expected changes in lipid metabolism [15], whereas carvedilol produced the opposite effects and a more favorable lipid profile. In this context, the availability of lipoprotein lipase for triglyceride hydrolysis would be reduced during atenolol therapy and enhanced during carvedilol therapy.

Our data also show that neither drug prolongs hypoglycemia and glucose recovery or interferes with glucagon and epinephrine responses after the insulin test. These findings have previously been seen during treatment with selective β (₁-adrenergic) antagonists [16], but no data on this topic were available for carvedilol. Among diabetic patients treated with insulin or oral antidiabetic drugs, no association has been found between an increased risk for hospital admission due to hypoglycemia and current use of β-blockers [17]. Diabetes mellitus and hypertension have been shown to be states of increased free-radical activity, which oxidatively stresses or injures the endothelium [18, 19]. Conjugated dienes and lipid peroxides are by-products of the lipid peroxidation of cellular structures induced by free radicals and can be conveniently measured as thiobarbituric-acid-reactive substances. Carvedilol, but not atenolol, significantly reduced levels of thiobarbituric-acid-reactive substances. These results agree with the results of in vitro studies indicating that carvedilol has the ability (an ability not shared by other β-blocking agents) to reduce neutrophil-mediated endothelial injury and inhibit copper ion-induced auto-oxidation of human low-density lipoprotein in vitro [6].

Although carvedilol seems to have potential advantages for the treatment of the diabetic patient with hypertension, we still do not know whether the beneficial biochemical changes associated with its use will improve the poor cardiovascular outlook of these patients. Moreover, some caution is necessary with the use of multiple comparisons; this makes confidence in our P values less secure. However, long-term β-blocker therapy seems to be associated with improved long-term survival in patients with non–insulin-dependent diabetes mellitus and coronary artery disease [20]. In this context, labetalol, the other mixed α- and β-blocking agent used in North America, has a β-blocking potency that is 50% lower than that of carvedilol [5]. Although further studies are needed to document the long-term effects of carvedilol with regard to benefit and lack of serious side effects, our results may help clinicians overcome their reluctance to prescribe β-blockers to patients who have both diabetes and hypertension.

From the Second University of Naples, Naples, Italy.

Drs. Acampora and De Angelis: Centro Antidiabete Lepanto, via Lepanto, 80100 Naples, Italy.

Dr. Ragone: Department of Biochemistry and Biophysics, Second University of Naples, 80138 Naples, Italy.