BLOOD PRESSURE
1992; 1; 92-101
Metabolic Effects of Pindolol and Propranolol in a Double-Blind Cross-Over Study in Hypertensive Patients H. LITHELL, T. POLLARE and B. VESSBY
Blood Press Downloaded from informahealthcare.com by University of Toronto on 12/25/14 For personal use only.
LitheU H, PoUare T, Vessby B. Metabolic effects ofpindolol and propranolol in a double-blind cross-over study in hypertensive patients. Blood Pressure 1992; 1: 92-101. Metabolic effects of pindolol and propranolol were investigated in a randomised study of double-blind, double-dummy design in 39 Caucasians with newly detected hypertension. Each active treatment period was 6 months long. A euglycaemic hyperinsulinaemic clamp test was done to measure insulin sensitivity, and i.v. glucose tolerance was investigated with insulin determinations. Lipoprotein concentrations were quantified and lipoprotein lipase activities were determined in muscle and adipose tissue and in plasma after heparin injection. The blood pressure was significantly reduced by both regimes. The insulin sensitivity index was decreased by 34% during propranolol treatment and by 17% during pindolol treatment. The insulin concentrations in plasma were elevated at the end of the i.v. glucose tolerance test but were not high enough to compensate for the insulin resistance, so HbAI, and glucose concentrations were increased. A significant reduction of lipoprotein lipase activity in skeletal muscle during propranolol treatment probably explains the pronounced increase in serum triglyceride concentration during propranolol treatment despite lower free fatty acids and higher lipoprotein lipase activity in adipose tissue. These changes of lipoprotein lipase activity were not correlated to the changes in insulin sensitivity. In summary, the metabolic effects were significantlyless pronounced with pindolol than with propranolol, which probably can be ascribed to the agonistic effect of pindolol on beta2 adrenoceptors. Key words: insulin, insulin sensitivity, pindolol.propranolo1, lipoproteins. lipoprotein lipase.
INTRODUCTION Since the antihypertensive effect of propranolol was first described by Prichard & Gilliam in 1964[I], a large number of beta-adrenoceptor-blocking drugs have been widely used in the treatment of hypertension [2]. The blood-pressure reduction is primarily related to a reduction in cardiac output which results in increased peripheral resistance [3]. This may give rise to subjective side-effects such as cold hands and feet. One way to alter the haemodynamic characteristics of a nonselective beta-adrenoceptor blocker, such as proprano101, is to alter the molecule so that it occupies and stimulates the beta2 receptor. This intrinsic symptomimic activity (ISA) is associated with vasodilation which is a characteristic feature of pindolol[4], one of several beta-adrenoceptor blocking drugs with ISA. Low density (LDL) as well as high density lipoprotein (HDL) cholesterol concentrations are strong and independent risk factors for coronary heart disease [5, 61. Changes, induced by lipid lowering drugs, of both LDL and of HDL cholesterol concentration have been shown to be associated with a changed risk of coronary heart disease [7, 81. Treatment of mild to moderate hypertension leads to reduced risk for coronary heart disease (CHD) but the risk is not reduced as much as could be expected from observational prospective studies [9-111. Some of this lack of effect may be attributable to the adverse effects on lipoproteins induced by treatment which counteracts the beneficial effect of blood-pressure lowering. A
metabolic angiopathic syndrome characterised by abdominal obesity, hypertension and deterioration of glucose and lipoprotein metabolism has been suggested to have a common denominator: insulin resistance [12141. The introduction of this concept has increased the understanding of the CHD risk factor concept and may further increase the understanding of the relative failure to reduce the risk of CHD when treating hypertension. It was recently demonstrated that treatment with thiazide diuretics [15] as well as with selective betablockers [16-181 is associated with an impairment of insulin resistance resulting in an elevation of insulin and glucose concentrations, and also in an elevation of serum triglycerides. In contrast, treatment with two drugs with vasodilating properties, prazosin [ 19-20] and captopril [15] have been associated with improvement in insulin sensitivity. The effects on insulin sensitivity and the entailing consequences on insulin concentrations may have implications for CHD risk not only via the relationship to hyperlipidaemia, hypertension and impaired glucose tolerance and diabetes but also because hyperinsulinaemia has been recognised as an independent risk factor for CHD [21-231 which may be related to the plaque formation induced by hyperinsulinaemia in in vitro studies [24]. In the present study we evaluated the effect of propranolol, a non-selective beta-blocking drug without ISA, and pindolol, a non-selective beta-blocking drug with a marked ISA, on insulin-mediated glucose disposal during long-term treatment and examined
Pindolol and propranolol in hypertensive patients
Blood Press Downloaded from informahealthcare.com by University of Toronto on 12/25/14 For personal use only.
their effects on glucose, insulin and lipid metabolism. Measurements of lipase activities were included and particular interest was directed to the question to what degree the lipoprotein changes were related to the changes in insulin resistance. MATERIALS AND METHODS Inclusion criteria All patients had previously untreated primary hypertension, defined as a stable resting supine diastolic pressure between 95 and 119 mmHg on at least two occasions during a 4- to 6-week single-blind placebo treatment period. All were newly detected hypertensive patients and had been followed for 3 to 4 months before entering the single-blind placebo trial. Exclusion criteria Patients with clinical or laboratory evidence of hepatic, renal, obstructive pulmonary disease, Raynaud‘s disease, or thyroid dysfunction were excluded. Patients were also excluded if they had a previous history of cardiovascular disease, major gastrointestinal surgery, renal impairment, diabetes mellitus (except diet-treated NIDDM), other endocrine disease, contraindications for treatment with beta-blocking agents and treatment with drugs for other diseases. Study design This was a randomised, double-blind, double-dummycontrolled, cross-over trial of 12 months’ duration, divided into four different treatment periods. The first treatment period was a single-blind placebo phase (4 to 6 weeks). The patients were given one propranolol placebo and one pindolol placebo twice daily, during this period. The second period (6 months) consisted of randomised treatment with fix dosages of either propranolol80 mg twice daily or pindolo15 mg twice daily (plus matching placebo twice daily). The third period consisted of a 4-week wash-out placebo period. During the fourth period (6 months) there was a change of active drugs. Measurements were done at the end of the run-in period and after each treatment period with active drug but not after the intervening wash-out period. Compliance with the drug treatment was assessed in all patients by interview and pill count. Subjective symptoms were actively evaluated by visual analogue scales (V.A.S.). For each of the 21 questions, there was a 100 mm (points) line. Zero points indicated no disturbing symptoms and 100 points maximally intensive, disturbing symptoms. Ratings during drug treat-
93
ment were compared with ratings during placebo and between drugs. There were also questions, using a 4point scale, concerning physical activities during leisure time and at work. The patients’ dietary habits were also actively investigated. All patients entering the study were given a thorough physical examination by one physician (T.P.). All blood pressures were measured by the same two nurses. Metabolic investigations were done at the end of each active treatment period, in the morning, 10-14 h after their last dose of medication. The patients were asked to refrain from smoking and from using caffeinecontaining beverages before the visit and had been fasting overnight (1 1-12 h). The insulin response to an intravenous glucose tolerance test (IVGTT) and insulin sensitivity were evaluated on two separate days with at least two days in between. Each participant was instructed to adhere to a diet rich in carbohydrates (at least 150 g of carbohydrates daily) and to refrain from extreme physical exercise or inactivity for at least 5 days before the investigation. N o other dietary or exercise instructions were given during the study. Ethical aspects Informed consent was obtained from all subjects after the design, purpose and possible side-effects of the study had been fully explained. The study protocol was approved by the Human Ethics Committee of the Medical Faculty of Uppsala University. Blood pressure and heart rate Measurements were made in the right arm. Systolic and diastolic blood pressure were defined as phases I and V, respectively, according to Korotkof. The blood pressure was measured (to the nearest two mm) with a mercury sphygmomanometer, three times in the supine position after 10min of rest and twice after standing for 1 min. The mean of the measurements, respectively, was used. A large cuff was used when appropriate. The heart rate was recorded before each blood pressure measurement. Metabolic investigations An intravenous glucose tolerance test was performed by injecting 300 mg glucose/kg body weight (b.w.) Plasma glucose concentration was determined by the glucose oxidase method (Optimate (TM), Ames-Gilford, Elkhart, IN, USA). The rate of disappearance of glucose was expressed as a k value calculated from the formula k = 100 x “og 2/Ti. The time (minutes) required for the glucose concentration to be halved was determined from the best fit of the measured values on
Blood Press Downloaded from informahealthcare.com by University of Toronto on 12/25/14 For personal use only.
94
H. Lithe11 et al.
semilogarithmic paper [25]. Immunoreactive insulin in plasma was assayed by a commercial radioimmunoassay kit (PhadesephnM) insulin radioimmunoassay, Pharmacia, Uppsala, Sweden). The peak insulin response was defined as the mean of the values obtained at 2, 4 and 6 minutes. The average fasting plasma insulin concentration was calculated from the values in four samples taken on two separate days. Glycated hemoglobin concentration was measured by fast performance liquid chromatography [26]. Lipoproteins were analysed by ultracentrifugation at a density of 1.006 to separate very low density lipoproteins (VLDL) and at 1.120 to separate HDL3 and precipitation (with phosphotungstate and magnesium chloride) f271. Triglyceride and cholesterol concentrations were measured by enzymatic techniques (Boehringer Mannheim, Germany) with a Multistat I11 I/LS centrifugal analyser (Instrumentation Laboratories, Lexington, MA, USA). The fasting fatty acid composition of the plasma cholesterol esters was determined by gas liquid chromatography as earlier described in [28]. Lipoprotein lipase activity was measured according to Lithell et al. (see [57]). Insulin sensitivity studies A two-hour euglycemic hyperinsulinemic clamp technique was used to estimate the sensitivity of patients to insulin while they were taking placebo and at the end of each treatment period [29]. The technique has been described in detail in [16]. The rate of infusion of insulin was 56 mU/m2 per min in all subjects, resulting in a mean plasma insulin concentration of 106 mU/1 (range 79-129 mU/l). The mean (SD) steady-state plasma glucose concentration during the clamp in patients taking placebo was 5.1 (0.2) mmol/l, and there was no significant change in concentration during different treatment periods. The coefficient of variation for the steady-state plasma glucose concentration for a single clamp was less than 4.5% on all occasions (on the average 3.6 (0.7%0)), and there were no significant changes in the coefficient of variation between different treatment periods. The amount of glucose taken up (mg-kg bw-' smin- ') during each clamp study was calculated for each 20-min interval after the first 20 min. The mean rate of glucose uptake for the last 60 min of the clamp was used as the main target variables. The index of sensitivity to insulin, a measure of tissue sensitivity to insulin expressed per unit of insulin, was calculated by dividing the amount of glucose taken up by the mean insulin concentration attained during the same period of the clamp [29]. The insulin concentrations attained suffiduring the insulin infusion (about 106 mU/1) is -- ----I -I
cient to suppress production of glucose by the liver in hypertensive patients with insulin resistance [30-351. Urinary losses of glucose were neglible under euglycaemic conditions. All other tests were carried out in the Department of Clinical Chemistry of the University Hospital, Uppsala. Body mass index (BMI) was calculated as weight (kg) divided by squared height (in metres). Statistical analysis An analysis of variance model with factors for treatment, time and patient was used. As the data were unbalanced because one subject dropped out and some values were missing, the mean values were not suitable for comparison. The results are presented as least square means [36] because they form the basis of the tests and estimates in the analysis and take the imbalance into account. A general problem with cross-over designs is the possible occurrence of carry-over effects, i.e. an influence from the first treatment period to the second. It is possible to test for such effects in this model. In this study no carry-over effects were detected and thus the results of both periods were combined. Comparisons were made of the two drugs' effects against each other and also for each drug against the results from the end of the run-in placebo period.
RESULTS Forty patients met the initial entry criteria and were enrolled in the placebo run-in. One of the patients failed to qualify during the run-in. Thus, 39 patients were metabolically characterised and continued on active drug (Table I). A pill count showed good patient compliance ( > 95% of tablets consumed by all subjects) and there was no change in dietary regimen or physical activities during the study according to active investigation. There were no significant carry-over effects between the drugs.
Table I. Anthropometric characteristics of the previously untreated hypertensive patients (n=39, 23 men, 16 women). Means k SD
Age (years) Height (cm) Weight (kg) Body mass index Waist-hip ratio Smokers
Men
Women
59.0f8.9 173.0f5.7 83.7f 11.6 28.0 & 3.5 0.93 f0.06 7
60.7 6.9 164.2k6.1 78.6 f22.5 29.0f 6.9 0.89 k0.08 3
Pindolol and propranolol in hypertensive patients Blood pressure
The mean blood pressure, both the supine and standing, was reduced significantly and equally well during treatment with each of the drugs (Table 11). During propranolol treatment the reduction in supine blood pressure was 22/14 mm Hg and during pindolol treatment 22/ 13. Propranolol and pindolol significantly decreased the heart rate in both the supine and standing positions. The reduction in heart rate was significantly smaller during pindolol than during propranolol treatment.
Blood Press Downloaded from informahealthcare.com by University of Toronto on 12/25/14 For personal use only.
Insulin sensitivity and glucose uptake
The insulin-mediated glucose uptake was significantly lower during treatment with propranolol and pindolol than during placebo, in spite of larger areas under the insulin curves during the clamp study by 10.8% (p=0.0133) for propranolol and by 6.8% ( p = 0 . 1 1 8 7 ) for propranolol, which in itself leads to a higher glucose uptake. The mean glucose uptake (during the last 60 min of the clamp) during treatment with placebo was 6.3 mg-kg bw-' min-' ( p < O . O O O l ) and decreased by 1.64 during treatment with propranolol and by 0.9 mg-kg bw-'-min-' ( p < 0.01) during treatment with pindolol. The difference between the two drugs was significant ( p < 0.01). The insulin sensitivity index is a measure of insulin sensitivity after correcting for the differences in prevailing insulin concentrations during the clamp. Insulin sensitivity index decreased from 6.5 to 4.3 (mg-kg bw-' -min-'.IU insulin-'. 100) when patients were taking propranolol ( p < O . O O O l ) and from 6.5 to 5.4 when patients were taking pindolol ( p < 0.01). The difference between the two drugs was significant ( p < 0.0 1).
Fasting insulin and glucose concentrations and response to intravenous glucose injection
Table 111and Figs 1A and B and 2A and B show data on fasting plasma insulin concentrations and related variables and the responses of plasma insulin and glucose concentrations during the intravenous glucose tolerance test before and after treatment with propranolol and pindolol. Fasting plasma insulin and glucose concentrations were significantly increased during propranolol but not during pindolol treatment, but the insulin values at the end of the IVGTT were higher during both treatments compared with placebo run-in values. The peak insulin value was increased from 59 mU/1 to 66 (p=O.O75) during pindolol treatment but remained unchanged during propranolol treatment. The k value for the disappearance of glucose during the glucose tolerance test decreased during both treatment regimens. There was an increase in HbAI, value during treatment with propranolol, with a significant difference between drugs. There was also a significant weight gain, about 1.5 kg, during treatment with both drugs (Table 111). Serum lipid and lipoprotein concentrations
VLDL triglyceride and cholesterol concentrations increased significantly during treatment with propranolol (Table IV), as did LDL triglyceride concentration, whereas treatment effects of pindolol did not differ from the placebo run-in results. The effect on LDL cholesterol concentration differed significantly between the drugs. HDL cholesterol concentration decreased during both drug regimens, significantly more so during propranolol than during pindolol treatment, reflecting reductions in HDL2 cholesterol concentration only. The ratio of LDL to HDL cholesterol increased by 12.4% ( p < 0.001) during treatment with propranolol
Table 11. Means and diferences f o r systolic ( S B P ) , diastolic ( D B P ) bloodpressure ( m m H g ) and heart rate ( H R , beatslmin) in the supine andstanding positions aregiven as the means at baseline (placebo; P ) andfinally the overaN estimated treatment eflects of propranolol ( A ) and pindolol ( B ) are presented
Variable SBP supine SBP standing DBP supine DBP standing HR supine HR standing
During treatment with placebo ( n = 39) 177
170 105 108
70 75
Effect of treatment with propranolol (n = 3 8 )
95
Effect of treatment with pindolol (n = 39)
-22*** -2o*** - 14*** - 13*** - 12*** - 12***
The degree of statistical significance is indicated by * p < 0.05, * * p < 0.01, ***p< 0.001.
Total difference in treatment effect 0-B) 0 (p=0.9146 -2 (p=0.4013) - 1 (p=0.3262) - 1 (p=0.4577) -7 (p=O.O001)
-6 (p=O.OOOl)
96
H . Lithell et al. ( B)
20
70 60-
= $15
g50E 40-
E E
.--3
Y
0) v)
8 ; 10
- 30-
Blood Press Downloaded from informahealthcare.com by University of Toronto on 12/25/14 For personal use only.
20-
I uPlacebo I I -w- Propranoiol I
* *
O
5
I
-10 0 10 20 30 40 50 60 70 80 90 30 40 50 60 70 80 90 Time (min) Time (min) Fig. 1. Plasma insulin and glucose concentrations during intravenous glucose tolerance test in 39 hypertensive patients receiving propranolol. *=p
1992; 1; 92-101
Metabolic Effects of Pindolol and Propranolol in a Double-Blind Cross-Over Study in Hypertensive Patients H. LITHELL, T. POLLARE and B. VESSBY
Blood Press Downloaded from informahealthcare.com by University of Toronto on 12/25/14 For personal use only.
LitheU H, PoUare T, Vessby B. Metabolic effects ofpindolol and propranolol in a double-blind cross-over study in hypertensive patients. Blood Pressure 1992; 1: 92-101. Metabolic effects of pindolol and propranolol were investigated in a randomised study of double-blind, double-dummy design in 39 Caucasians with newly detected hypertension. Each active treatment period was 6 months long. A euglycaemic hyperinsulinaemic clamp test was done to measure insulin sensitivity, and i.v. glucose tolerance was investigated with insulin determinations. Lipoprotein concentrations were quantified and lipoprotein lipase activities were determined in muscle and adipose tissue and in plasma after heparin injection. The blood pressure was significantly reduced by both regimes. The insulin sensitivity index was decreased by 34% during propranolol treatment and by 17% during pindolol treatment. The insulin concentrations in plasma were elevated at the end of the i.v. glucose tolerance test but were not high enough to compensate for the insulin resistance, so HbAI, and glucose concentrations were increased. A significant reduction of lipoprotein lipase activity in skeletal muscle during propranolol treatment probably explains the pronounced increase in serum triglyceride concentration during propranolol treatment despite lower free fatty acids and higher lipoprotein lipase activity in adipose tissue. These changes of lipoprotein lipase activity were not correlated to the changes in insulin sensitivity. In summary, the metabolic effects were significantlyless pronounced with pindolol than with propranolol, which probably can be ascribed to the agonistic effect of pindolol on beta2 adrenoceptors. Key words: insulin, insulin sensitivity, pindolol.propranolo1, lipoproteins. lipoprotein lipase.
INTRODUCTION Since the antihypertensive effect of propranolol was first described by Prichard & Gilliam in 1964[I], a large number of beta-adrenoceptor-blocking drugs have been widely used in the treatment of hypertension [2]. The blood-pressure reduction is primarily related to a reduction in cardiac output which results in increased peripheral resistance [3]. This may give rise to subjective side-effects such as cold hands and feet. One way to alter the haemodynamic characteristics of a nonselective beta-adrenoceptor blocker, such as proprano101, is to alter the molecule so that it occupies and stimulates the beta2 receptor. This intrinsic symptomimic activity (ISA) is associated with vasodilation which is a characteristic feature of pindolol[4], one of several beta-adrenoceptor blocking drugs with ISA. Low density (LDL) as well as high density lipoprotein (HDL) cholesterol concentrations are strong and independent risk factors for coronary heart disease [5, 61. Changes, induced by lipid lowering drugs, of both LDL and of HDL cholesterol concentration have been shown to be associated with a changed risk of coronary heart disease [7, 81. Treatment of mild to moderate hypertension leads to reduced risk for coronary heart disease (CHD) but the risk is not reduced as much as could be expected from observational prospective studies [9-111. Some of this lack of effect may be attributable to the adverse effects on lipoproteins induced by treatment which counteracts the beneficial effect of blood-pressure lowering. A
metabolic angiopathic syndrome characterised by abdominal obesity, hypertension and deterioration of glucose and lipoprotein metabolism has been suggested to have a common denominator: insulin resistance [12141. The introduction of this concept has increased the understanding of the CHD risk factor concept and may further increase the understanding of the relative failure to reduce the risk of CHD when treating hypertension. It was recently demonstrated that treatment with thiazide diuretics [15] as well as with selective betablockers [16-181 is associated with an impairment of insulin resistance resulting in an elevation of insulin and glucose concentrations, and also in an elevation of serum triglycerides. In contrast, treatment with two drugs with vasodilating properties, prazosin [ 19-20] and captopril [15] have been associated with improvement in insulin sensitivity. The effects on insulin sensitivity and the entailing consequences on insulin concentrations may have implications for CHD risk not only via the relationship to hyperlipidaemia, hypertension and impaired glucose tolerance and diabetes but also because hyperinsulinaemia has been recognised as an independent risk factor for CHD [21-231 which may be related to the plaque formation induced by hyperinsulinaemia in in vitro studies [24]. In the present study we evaluated the effect of propranolol, a non-selective beta-blocking drug without ISA, and pindolol, a non-selective beta-blocking drug with a marked ISA, on insulin-mediated glucose disposal during long-term treatment and examined
Pindolol and propranolol in hypertensive patients
Blood Press Downloaded from informahealthcare.com by University of Toronto on 12/25/14 For personal use only.
their effects on glucose, insulin and lipid metabolism. Measurements of lipase activities were included and particular interest was directed to the question to what degree the lipoprotein changes were related to the changes in insulin resistance. MATERIALS AND METHODS Inclusion criteria All patients had previously untreated primary hypertension, defined as a stable resting supine diastolic pressure between 95 and 119 mmHg on at least two occasions during a 4- to 6-week single-blind placebo treatment period. All were newly detected hypertensive patients and had been followed for 3 to 4 months before entering the single-blind placebo trial. Exclusion criteria Patients with clinical or laboratory evidence of hepatic, renal, obstructive pulmonary disease, Raynaud‘s disease, or thyroid dysfunction were excluded. Patients were also excluded if they had a previous history of cardiovascular disease, major gastrointestinal surgery, renal impairment, diabetes mellitus (except diet-treated NIDDM), other endocrine disease, contraindications for treatment with beta-blocking agents and treatment with drugs for other diseases. Study design This was a randomised, double-blind, double-dummycontrolled, cross-over trial of 12 months’ duration, divided into four different treatment periods. The first treatment period was a single-blind placebo phase (4 to 6 weeks). The patients were given one propranolol placebo and one pindolol placebo twice daily, during this period. The second period (6 months) consisted of randomised treatment with fix dosages of either propranolol80 mg twice daily or pindolo15 mg twice daily (plus matching placebo twice daily). The third period consisted of a 4-week wash-out placebo period. During the fourth period (6 months) there was a change of active drugs. Measurements were done at the end of the run-in period and after each treatment period with active drug but not after the intervening wash-out period. Compliance with the drug treatment was assessed in all patients by interview and pill count. Subjective symptoms were actively evaluated by visual analogue scales (V.A.S.). For each of the 21 questions, there was a 100 mm (points) line. Zero points indicated no disturbing symptoms and 100 points maximally intensive, disturbing symptoms. Ratings during drug treat-
93
ment were compared with ratings during placebo and between drugs. There were also questions, using a 4point scale, concerning physical activities during leisure time and at work. The patients’ dietary habits were also actively investigated. All patients entering the study were given a thorough physical examination by one physician (T.P.). All blood pressures were measured by the same two nurses. Metabolic investigations were done at the end of each active treatment period, in the morning, 10-14 h after their last dose of medication. The patients were asked to refrain from smoking and from using caffeinecontaining beverages before the visit and had been fasting overnight (1 1-12 h). The insulin response to an intravenous glucose tolerance test (IVGTT) and insulin sensitivity were evaluated on two separate days with at least two days in between. Each participant was instructed to adhere to a diet rich in carbohydrates (at least 150 g of carbohydrates daily) and to refrain from extreme physical exercise or inactivity for at least 5 days before the investigation. N o other dietary or exercise instructions were given during the study. Ethical aspects Informed consent was obtained from all subjects after the design, purpose and possible side-effects of the study had been fully explained. The study protocol was approved by the Human Ethics Committee of the Medical Faculty of Uppsala University. Blood pressure and heart rate Measurements were made in the right arm. Systolic and diastolic blood pressure were defined as phases I and V, respectively, according to Korotkof. The blood pressure was measured (to the nearest two mm) with a mercury sphygmomanometer, three times in the supine position after 10min of rest and twice after standing for 1 min. The mean of the measurements, respectively, was used. A large cuff was used when appropriate. The heart rate was recorded before each blood pressure measurement. Metabolic investigations An intravenous glucose tolerance test was performed by injecting 300 mg glucose/kg body weight (b.w.) Plasma glucose concentration was determined by the glucose oxidase method (Optimate (TM), Ames-Gilford, Elkhart, IN, USA). The rate of disappearance of glucose was expressed as a k value calculated from the formula k = 100 x “og 2/Ti. The time (minutes) required for the glucose concentration to be halved was determined from the best fit of the measured values on
Blood Press Downloaded from informahealthcare.com by University of Toronto on 12/25/14 For personal use only.
94
H. Lithe11 et al.
semilogarithmic paper [25]. Immunoreactive insulin in plasma was assayed by a commercial radioimmunoassay kit (PhadesephnM) insulin radioimmunoassay, Pharmacia, Uppsala, Sweden). The peak insulin response was defined as the mean of the values obtained at 2, 4 and 6 minutes. The average fasting plasma insulin concentration was calculated from the values in four samples taken on two separate days. Glycated hemoglobin concentration was measured by fast performance liquid chromatography [26]. Lipoproteins were analysed by ultracentrifugation at a density of 1.006 to separate very low density lipoproteins (VLDL) and at 1.120 to separate HDL3 and precipitation (with phosphotungstate and magnesium chloride) f271. Triglyceride and cholesterol concentrations were measured by enzymatic techniques (Boehringer Mannheim, Germany) with a Multistat I11 I/LS centrifugal analyser (Instrumentation Laboratories, Lexington, MA, USA). The fasting fatty acid composition of the plasma cholesterol esters was determined by gas liquid chromatography as earlier described in [28]. Lipoprotein lipase activity was measured according to Lithell et al. (see [57]). Insulin sensitivity studies A two-hour euglycemic hyperinsulinemic clamp technique was used to estimate the sensitivity of patients to insulin while they were taking placebo and at the end of each treatment period [29]. The technique has been described in detail in [16]. The rate of infusion of insulin was 56 mU/m2 per min in all subjects, resulting in a mean plasma insulin concentration of 106 mU/1 (range 79-129 mU/l). The mean (SD) steady-state plasma glucose concentration during the clamp in patients taking placebo was 5.1 (0.2) mmol/l, and there was no significant change in concentration during different treatment periods. The coefficient of variation for the steady-state plasma glucose concentration for a single clamp was less than 4.5% on all occasions (on the average 3.6 (0.7%0)), and there were no significant changes in the coefficient of variation between different treatment periods. The amount of glucose taken up (mg-kg bw-' smin- ') during each clamp study was calculated for each 20-min interval after the first 20 min. The mean rate of glucose uptake for the last 60 min of the clamp was used as the main target variables. The index of sensitivity to insulin, a measure of tissue sensitivity to insulin expressed per unit of insulin, was calculated by dividing the amount of glucose taken up by the mean insulin concentration attained during the same period of the clamp [29]. The insulin concentrations attained suffiduring the insulin infusion (about 106 mU/1) is -- ----I -I
cient to suppress production of glucose by the liver in hypertensive patients with insulin resistance [30-351. Urinary losses of glucose were neglible under euglycaemic conditions. All other tests were carried out in the Department of Clinical Chemistry of the University Hospital, Uppsala. Body mass index (BMI) was calculated as weight (kg) divided by squared height (in metres). Statistical analysis An analysis of variance model with factors for treatment, time and patient was used. As the data were unbalanced because one subject dropped out and some values were missing, the mean values were not suitable for comparison. The results are presented as least square means [36] because they form the basis of the tests and estimates in the analysis and take the imbalance into account. A general problem with cross-over designs is the possible occurrence of carry-over effects, i.e. an influence from the first treatment period to the second. It is possible to test for such effects in this model. In this study no carry-over effects were detected and thus the results of both periods were combined. Comparisons were made of the two drugs' effects against each other and also for each drug against the results from the end of the run-in placebo period.
RESULTS Forty patients met the initial entry criteria and were enrolled in the placebo run-in. One of the patients failed to qualify during the run-in. Thus, 39 patients were metabolically characterised and continued on active drug (Table I). A pill count showed good patient compliance ( > 95% of tablets consumed by all subjects) and there was no change in dietary regimen or physical activities during the study according to active investigation. There were no significant carry-over effects between the drugs.
Table I. Anthropometric characteristics of the previously untreated hypertensive patients (n=39, 23 men, 16 women). Means k SD
Age (years) Height (cm) Weight (kg) Body mass index Waist-hip ratio Smokers
Men
Women
59.0f8.9 173.0f5.7 83.7f 11.6 28.0 & 3.5 0.93 f0.06 7
60.7 6.9 164.2k6.1 78.6 f22.5 29.0f 6.9 0.89 k0.08 3
Pindolol and propranolol in hypertensive patients Blood pressure
The mean blood pressure, both the supine and standing, was reduced significantly and equally well during treatment with each of the drugs (Table 11). During propranolol treatment the reduction in supine blood pressure was 22/14 mm Hg and during pindolol treatment 22/ 13. Propranolol and pindolol significantly decreased the heart rate in both the supine and standing positions. The reduction in heart rate was significantly smaller during pindolol than during propranolol treatment.
Blood Press Downloaded from informahealthcare.com by University of Toronto on 12/25/14 For personal use only.
Insulin sensitivity and glucose uptake
The insulin-mediated glucose uptake was significantly lower during treatment with propranolol and pindolol than during placebo, in spite of larger areas under the insulin curves during the clamp study by 10.8% (p=0.0133) for propranolol and by 6.8% ( p = 0 . 1 1 8 7 ) for propranolol, which in itself leads to a higher glucose uptake. The mean glucose uptake (during the last 60 min of the clamp) during treatment with placebo was 6.3 mg-kg bw-' min-' ( p < O . O O O l ) and decreased by 1.64 during treatment with propranolol and by 0.9 mg-kg bw-'-min-' ( p < 0.01) during treatment with pindolol. The difference between the two drugs was significant ( p < 0.01). The insulin sensitivity index is a measure of insulin sensitivity after correcting for the differences in prevailing insulin concentrations during the clamp. Insulin sensitivity index decreased from 6.5 to 4.3 (mg-kg bw-' -min-'.IU insulin-'. 100) when patients were taking propranolol ( p < O . O O O l ) and from 6.5 to 5.4 when patients were taking pindolol ( p < 0.01). The difference between the two drugs was significant ( p < 0.0 1).
Fasting insulin and glucose concentrations and response to intravenous glucose injection
Table 111and Figs 1A and B and 2A and B show data on fasting plasma insulin concentrations and related variables and the responses of plasma insulin and glucose concentrations during the intravenous glucose tolerance test before and after treatment with propranolol and pindolol. Fasting plasma insulin and glucose concentrations were significantly increased during propranolol but not during pindolol treatment, but the insulin values at the end of the IVGTT were higher during both treatments compared with placebo run-in values. The peak insulin value was increased from 59 mU/1 to 66 (p=O.O75) during pindolol treatment but remained unchanged during propranolol treatment. The k value for the disappearance of glucose during the glucose tolerance test decreased during both treatment regimens. There was an increase in HbAI, value during treatment with propranolol, with a significant difference between drugs. There was also a significant weight gain, about 1.5 kg, during treatment with both drugs (Table 111). Serum lipid and lipoprotein concentrations
VLDL triglyceride and cholesterol concentrations increased significantly during treatment with propranolol (Table IV), as did LDL triglyceride concentration, whereas treatment effects of pindolol did not differ from the placebo run-in results. The effect on LDL cholesterol concentration differed significantly between the drugs. HDL cholesterol concentration decreased during both drug regimens, significantly more so during propranolol than during pindolol treatment, reflecting reductions in HDL2 cholesterol concentration only. The ratio of LDL to HDL cholesterol increased by 12.4% ( p < 0.001) during treatment with propranolol
Table 11. Means and diferences f o r systolic ( S B P ) , diastolic ( D B P ) bloodpressure ( m m H g ) and heart rate ( H R , beatslmin) in the supine andstanding positions aregiven as the means at baseline (placebo; P ) andfinally the overaN estimated treatment eflects of propranolol ( A ) and pindolol ( B ) are presented
Variable SBP supine SBP standing DBP supine DBP standing HR supine HR standing
During treatment with placebo ( n = 39) 177
170 105 108
70 75
Effect of treatment with propranolol (n = 3 8 )
95
Effect of treatment with pindolol (n = 39)
-22*** -2o*** - 14*** - 13*** - 12*** - 12***
The degree of statistical significance is indicated by * p < 0.05, * * p < 0.01, ***p< 0.001.
Total difference in treatment effect 0-B) 0 (p=0.9146 -2 (p=0.4013) - 1 (p=0.3262) - 1 (p=0.4577) -7 (p=O.O001)
-6 (p=O.OOOl)
96
H . Lithell et al. ( B)
20
70 60-
= $15
g50E 40-
E E
.--3
Y
0) v)
8 ; 10
- 30-
Blood Press Downloaded from informahealthcare.com by University of Toronto on 12/25/14 For personal use only.
20-
I uPlacebo I I -w- Propranoiol I
* *
O
5
I
-10 0 10 20 30 40 50 60 70 80 90 30 40 50 60 70 80 90 Time (min) Time (min) Fig. 1. Plasma insulin and glucose concentrations during intravenous glucose tolerance test in 39 hypertensive patients receiving propranolol. *=p
