Cardiovascular Drags and Therapy 1992;6:219-223 © Kluwer Academic Publishers, Boston. Printed in U.S.A.
Effects of Hydrochlorothiazide, Amiloride, and Lisinopril on the Metabolic Response to Adrenaline Infusions in Normal Subjects Ole Hansen, Bengt W Johansson Cardiology Section, General Hospital, MalmiJ, Sweden
Summary. Twelve healthy male volunteers were given adrenaline infusions, 0.05 ~tg/kg body weight/min over 120 minutes in order to achieve serum adrenaline concentrations comparable with those seen in acute myocardial infarction. The infusions were given on four occasions, at intervals of at least 4 weeks. Before the infusions the subjects were given, in random order, 14 days of p r e t r e a t m e n t with placebo, hydrochlorothiazide 50 mg once daily, amiloride 10 mg once daily, or lisinorpil 20 mg once daily. The adrenaline infusion induced a drop in serum potassium of the same magnitude in all four groups, with the lowest absolute value after hydrochlorothiazide because of the lowest pre-adrenaline level. The infusioninduced decreases in serum calcium and magnesium were of the same magnitude in all groups, with the absolute calcium being least low in the hydrochlorothiazide group because of the highest preinfusion value. Preinfusion serum urate was highest after hydrochlorothiazide and fell during the adrenaline infusion in all groups, although not significantly. Blood glucose increased during the adrenaline infusion in all groups, but significantly more after hydrochlorothiazide and amiloride t h a n after lisinopril. Heart rate increased during the adrenaline infusion in all groups but least after lisinopril. QTc preinfusion was longer after hydrochlorothiazide t h a n after amiloride and placebo, but the infusion-induced prolongation of QTc was of the same magnitude in all pretreatment groups. Since our results were obtained in short-term experiments in normal subjects, t h e i r clinical relevance is questionable, but they support the view t h a t ACE inhibitors may have certain metabolic advantages over diuretics. Cardiovasc Drugs Ther 1992;6:219-223 Key Words. adrenaline, hypertension, metabolism, amiloride, hydrochlorothiazide, lisinopril, electrocardiogram
Recently there has been much discussion of the longterm effects of the metabolic changes (e.g., raised blood sugar with thiazides) seen with some of the drugs used in the treatment of heart failure and hypertension. In this paper we look at another aspect of the metabolic effects of these drugs, namely, their effects on the metabolic response to adrenaline infusions in normal subjects. Myocardial infarction, a not uncommon event in patients with heart failure and hypertension, is associated with elevated plasma adrenaline levels and with metabolic and ionic changes. These changes may play a role in the malig-
nant arrhythmias sometimes seen with myocardial ischemia and infarction [1,2]. We have previously shown that adrenaline infusions in healthy volunteers [1,3,4] cause metabolic and ionic changes similar to those seen in myocardial infarction, and in the present study we use the same model to investigate whether pretreatment with some of the drugs used in the treatment of heart failure and hypertension has any effect on these changes. We have also measured some of the metabolic and ionic changes induced by these drugs during the 2-week pretreatment period prior to the adrenaline infusion. The drugs we have studied are hydrochlorothiazide (a thiazide, a class of drugs that for years has been first-choice therapy for hypertension and heart failure), lisinopril (an ACE inhibitor, a class of drugs that many think should now be first-choice therapy), and amiloride (a potassium-sparing diuretic). Although amiloride on its own is not a universally accepted therapy, it is used in several centers; also, it is often used in combination with a thiazide, and therefore we decided to study it. In previous studies we have investigated beta blockers and calcium antagonists [1,3].
Patients and Methods Twelve healthy male volunteers, aged 23-38 (mean 28.8) years, receiving no regular medication, were each given, on four separate occasions, each separated by at least 4 weeks, an adrenaline infusion at a rate of 0.05 ~g/min/kg body weight over 120 minutes. The adrenaline infusion was given via a cannulated antecubital vein, with the volunteer at rest in the supine position. A cannula was inserted in the opposite upper limb for blood sampling. Before the adrenaline infusion the volunteers were pretreated for 14 days with placebo, hydrochlorothiazide 50 mg once daily, amiloride 10 mg once daily, or lisinopril 20 mg once daily in a randomized order, with each of the volunteers
Address for correspondence and reprint requests: Bengt W. Johansson, MD, Cardiology Section, General Hospital, S-21401 Malm6, Sweden.
219
220
Hansen and Johansson
receiving each of the four pretreatments (Latin square design). On each occasion the adrenaline infusion was started at 1 p.m., and approximately 5 hours before the infusion the volunteer was given a tablet of the current pretreatment. Fifteen minutes before the adrenaline infusion, at the start of the infusion, every 15 minutes during the infusion, and 15 and 30 minutes after the infusion, blood was drawn for analysis of serum potassium, magnesium, calcium, phosphate, urate, albumin, creatinine, and blood glucose using the routine methods of a local laboratory. An electrocardiogram (leads I, II, III, aVR, aVL, and aVF) was recorded continuously from the start of the adrenaline infusion until 30 minutes after the infusion at a speed of l0 mm/sec, which was increased to 50 mm/sec at the times of blood sampling. The electrocardiogram was analyzed for heart rate, T-wave amplitude, and QRS and QT durations. QTc duration was calculated by the use of Bazett's formula. QT was measured from the beginning of the QRS complex to the end of the T wave, which was defined as the point where the descending limb of the T wave returned to the isoelectric line, defined by the PQ segment. For the calculations, the average of three representative beats in the lead with the longest QT segment during sinus rhythm was used. T-wave amplitude was measured from the isoelectric line to the highest point of the T wave in lead II, in which the duration of the QRS complex was also measured. The average of three representative cardiac cycles during sinus rhythm was used for the calculations. Just after blood sampling, systolic blood pressure and diastolic blood pressure were determined by the cuff method (phase 5 for diastolic blood pressure). The results are expressed as the mean - standard deviation. For statistical analysis a multivariate analysis of variance for repeated measures design (MANOVA) was used with time during the adrenaline infusion, and the various pretreatments were the within-subject factors [5]. If the analysis of variance suggested differences in baseline values between the various pretreatments, these differences were evaluated by t tests for paired samples. All statistical tests were two sided, p < 0.05 was considered to be statistically significant. The protocol was approved by the local ethics committee, and all volunteers gave informed consent before they were included in the study.
Results
Metabolic results The effects of an adrenaline infusion on the various metabolic variables measured in the present study are summarized in Table 1. The adrenaline infusion had significant effects on all variables except urate, causing rapid falls in S-potassium and S-phosphate, which
were apparent after 15 minutes, whereas the drops in S-calcium and S-magnesium occurred more slowly. A marked increase in the B-glucose was also seen, as well as minor, although still statistically significant, changes in S-albumin and S-creatinine. Metabolic effects of the pretreatment before adrenaline infusion When the effects of the various pretreatments on the metabolic variables before the start of the adrenaline infusion were compared, hydrochlorothiazide was found to lower S-potassium significantly (p < 0.01) compared with placebo, and to increase significantly S-albumin (p < 0.01) compared with placebo and Surate (p < 0.001) compared with all the other pretreatments. Pretreatments with amiloride or lisinopril were not associated with any significant effects on the metabolic variables at baseline when compared with placebo. Effects of the active pretreatments on the metabolic response to adrenaline infusion A significant interaction was found between the effect of the adrenaline infusion and the various pretreatments on B-glucose. Thus, pretreatment with hydrochlorothiazide caused a greater increase in B-glucose during the adrenaline infusion than that seen with placebo (p < 0.044), with a value of 9.44 -+ 2.5 mmol/1 at time 120 minutes. Also, pretreatment with amiloride was associated with higher B-glucose values during the adrenaline infusion than those seen with placebo, but the overall response was not statistically significant compared with placebo. After pretreatment with lisinopril, B-glucose rose less during the adrenaline infusion (B-glucose = 8.23 - 1.6 mmol/1 at time 120 minutes) than after the other pretreatments, although the difference was not statistically significant when compared with placebo. However, when lisinopril was compared with the diuretics, it was found that both hydrochlorothiazide (p < 0.001) and amiloride (p < 0.01) accentuated the adrenaline-induced increase in B-glucose. Lisinopril pretreatment also significantly modified the adrenaline-induced changes in S-phosphate (p < 0.001). Thus, after lisinopril the fall in S-phosphate during the adrenaline infusion was smaller than that seen after placebo or hydrochlorothiazide. The adrenaline-induced falls in S-magnesium and S-potassium were similar after all pretreatments, but due to its potassium-lowering effect before the start of the adrenaline infusion, the lowest S-potassium (2.98 -+ 0.37 mmol/1) during the infusion was seen after pretreatment with hydrochlorothiazide. Hemodynamic and electrocardiographic results These results are summarized in table 2. Adrenaline caused significant increases in heart rate and systolic blood pressure, and a significant fall in diastolic blood
Adrenaline-Induced Electrolyte Changes After Antihypertensive Drags
221
Table 1. Changes in variables before and after 120-minute adrenaline infusion (0.05 #g/kg body weight) to male healthy volunteers after 14 day pretreatment with placebo, hydrochlorothiazide 50 mg daily, amiloride 10 mg once daily, and lisinopril 20 mg once daily in a randomized order
Serum potassium (mmol/1) Serum magnesium (mmol/1) Serum calcium (mmol/1) Serum phosphate (mmol/1) Serum urate (~mol/1) Serum creatinine (~mol/1) S e r u m albumin (g/l) Blood glucose (retool/l)
Before After 30 nun Before After 30 m m Before After 30 mm Before After 30 m m Before After 30 mm Before After 30 mm Before After 30 nun Before After 30 mln
Placebo
Hydrochlorothiazide
Amiloride
Lisinopril
4.03 3.39 3.63 0.85 0.78 0.76 2.38 2.31 2.27 1.16 0.92 1.05 350 349 351 105 112 108 40 40 38 4.6 8.6 7.4
3.51 2.98 3.23 0.85 0.79 0.79 2.44 2.33 2.33 1.16 0.82 0.97 428 419 423 106 118 113 42 40 39 4.5 9.4 7.6
4.13 3.52 3.79 0.84 0.77 0.78 2.40 2.28 2.29 1.12 0.96 1.05 344 340 343 104 113 111 40 39 38 4.6 9.4 8.1
4.14 3.53 3.77 0.84 0.78 0.79 2.39 2.31 2.29 1.08 0.96 1.07 361 353 358 100 112 105 40 39 39 4.6 8.2 7.0
-+ 0.21 -+ 0.31 -+ 0.22 -+ 0.07 -+ 0.05 -+ 0.07 -4- 0.07 -+ 0.11 -+ 0.10 -+ 0.16 -+ 0.15 -+ 0.17 -+ 44 -+ 52 -+ 53 -+ 20 -+ 16 -+ 15 -+ 1.9 -+ 2.5 -+ 2.7 -+ 0.68 -+ 1.88 -+ 1.62
-+ 0.33 -+ 0.37 -+ 0.37 -+ 0.08 -+ 0.08 -+ 0.09 -+ 0.08 -+ 0.11 -+ 0.10 -+ 0.15 -+ 0.16 -+ 0.17 -+ 56 -+ 61 - 58 -+ 13 -+ 15 -+ 15 -+ 2.1 -+ 1.6 - 2.1 -+ 0.70 -+ 2.50 -+ 2.05
-+ 0.38 -+ 0.34 -+ 0.23 -+ 0.06 -+ 0.05 +- 0.06 -+ 0.11 -+ 0.08 -+ 0.09 -+ 0.16 -+ 0.15 -+ 0.17 +_ 40 -+ 39 -+ 37 -+ 8 -+ 13 + 13 + 1.8 - 1.5 -+ 1.7 -+ 0.76 -+ 1.62 -+ 1.58
-+ 0.23 -+ 0.24 -+ 0.24 -+ 0.06 -+ 0.06 -+ 0.08 - 0.11 - 0.09 -+ 0.09 + 0.14 -+ 0.13 -+ 0.11 + 52 -+ 49 +_ 50 -+ 10 -+ 12 -+ 11 -+ 2.7 - 2.5 -+ 2.7 -+ 0.89 + 1.60 -+ 1.11
Values are given as mean ± standard deviation just before the adrenaline infusion (before), after 120 minutes of infusion (after), and 30 minutes after the end of the infusion (30 min).
Table 2. Hemodynamic and electrocardiographic results
H e a r t rate (beats/rain) Systolic blood p r e s s u r e (mmHg) Diastolic blood p r e s s u r e (mmHg) Corrected QT duration (msec) T-wave amplitude m m (10 m m = QRS duration (msec)
1 mV)
Abbreviations as in Table 1.
Before After 30 min Before After 30 min Before After 30 min Before After 30 min Before After 30 min Before After 30 min
Placebo
Hydrochlorothiazide
Amiloride
Lisinopril
62-9 74-+ 14 68 -+ 12 130 -+ 11 143 -+ 7 133 -+ 7 79-+8 63-+8 78-+7 0.40 -+ 0.02 0.43 - 0.03 0.41 -+ 0.02 4-+1 3-+1 3-+1 0.090 -+ 0.007 0.095 -+ 0.007 0.091 -+ 0.009
66-+9 79-+7 70 -+ 10 129 -+ 9 141 -+ 8 130 -+ 8 79-+7 63-+9 80-+8 0.41 -+ 0.02 0.45 -+ 0.04 0.42 -+ 0.03 4-+1 3+1 3-+1 0.093 -+ 0.008 0.096 -+ 0.007 0.092 -+ 0.008
66-+ 12 78-+ 12 70 -+ 12 127 -+ 8 143 -+ 9 131 -+ 6 78-+8 66-+9 80-+8 0.40 -+ 0.02 0.43 - 0.02 0.41 -+ 0.02 4-+1 3-+1 4-+1 0.088 -+ 0.008 0.093 -+ 0.009 0.091 -+ 0.008
64-+ 13 72-+ 12 67 -+ 12 125 -+ 11 135 -+ 10 125 -+ 8 75-+6 63-+6 76-+5 0.40 -+ 0.02 0.43 -+ 0.02 0.41 -+ 0.02 5-+1 3-+1 4-+1 0.090 - 0.009 0.093 -4- 0.011 0.092 -+ 0.009
222
Hansen and Johansson
pressure. There were no statistically significant differences between the various pretreatments on the adrenaline effects of these parameters. During the adrenaline infusion, significant prolongations of the QRS and QTc durations, and a flattening of the T-wave amplitude, were seen, but the responses were not significantly modified by the various pretreatments. On the other hand, the QTc duration (0.413) at baseline before the adrenaline infusion was significantly prolonged (p < 0.01) after hydrochlorothiazide when compared with placebo. The QRS duration at baseline was 0.093 after hydrochlorothiazide, which was significantly longer (p < 0.01) than after amiloride (0.088), and was also significantly (p < 0.05) longer than the QRS duration (0.090) seen after pretreatment with lisinopril.
Discussion
Patients with heart failure and hypertension may be exposed to drugs for a long time. It is therefore important to consider even minor side effects in order to select the optimal therapy for the individual patient. This is especially true for patients with mild hypertension or heart failure, whereas patients with more severe disease may need the whole therapeutic armamentarium available and may have to tolerate side effects that are unacceptable in patients with milder disease. Recently there has been much discussion about the exact place of adrenergic beta-receptor blocking agents, and perhaps especially diuretics, in the treatment of hypertension. One reason for this is that we now think that hypertension is often not a disease per se but part of a metabolic syndrome and that components of this syndrome may be exacerbated by the known side effects of these drugs. In addition to an elevated blood pressure, this metabolic syndrome encompasses several factors, including obesity, disturbances in lipid metabolism (elevated serum triglycerides and VLDL cholesterol but low HDL cholesterol), disturbances in urate and calciummagnesium metabolism, and disturbances in carbohydrate metabolism [6]. The pathogenetic mechanisms behind the metabolic syndrome are still far from clear, but involvement of the sympathetic nervous system has been proposed [7]. We have used our adrenaline infusion model to imitate in a defined experimental situation the stress that all individuals, including those at a high risk of developing cardiovascular complications, are exposed to in their daily lives. We have chosen a circulating plasma adrenaline concentration of the magnitude known to occur in patients with an acute myocardial infarction [1]. Mental stress has been shown to increase venous plasma adrenaline levels, although to a lesser degree than acute infarction [8]. In this study, in agreement with our previous study, the adrenaline infusion caused decreases in se-
rum potassium, phosphate, magnesium, calcium, diastolic blood pressure, and T-wave amplitude, and increases in blood glucose, heart rate, systolic blood pressure, and QTc. An analysis of the myocardial infarction register in Malm5 has confirmed the results of several other studies and has shown that hypokalemia in the initial phase of an infarction is associated with an increased number of circulatory arrests due to ventricular fibrillation or tachycardia [9], the frequency of arrests being inversely related to the serum potassium value [10]. In patients at risk of an acute myocardial infarction, it seems mandatory to avoid hypokalemia. A low serum phosphate in primary hypertension was described early [11]. Increased serum urate values have been reported in hypertension, although this finding is doubtful when correction for obesity is performed [12]. Adrenaline elevated the blood glucose level [1]. Although the drugs tested in this study had different effects on the serum phosphate and urate changes, we do not feel enough is known of the consequences (if any) of these changes to speculate on the importance of the different effects of the drugs on them. The higher serum creatinine and serum albumin baseline values seen with hydrochlorothiazide can probably be ascribed to the slight dehydration induced by this drug. The differences in serum creatinine after the adrenaline infusion were small in all but the pretreatment groups and were related to changes as a result of the infusion. It is of interest that the serum albumin baseline value was higher after pretreatment with hydrochlorothiazide than after amiloride, suggesting that the diuretic effect of amiloride in the dosage used in small compared with its potassium-sparing effect. The relation between magnesium deficiency and ischemic heart disease is of great interest. We observed in a previous study that serum magnesium decreased during the initial phase of the infarction, although it still remained within the reference range of the laboratory [13], but in other studies a transient hypomagnesemia was seen. This agrees with the beneficial effect of magnesium infusions in patients with an acute myocardial infarction, as reported by Rasmussen et al. [14], but confirmation of this must await the result of larger studies, such as ISIS-4. Low levels of serum potassium, calcium, and magnesium are all known to have the potential to provoke malignant arrhythmias [2,9,13]. In a previous study we have shown that the decrease in these electrolytes induced by adrenaline is mediated by the beta2-receptors [3]. The liability to ventricular arrhythmias seen with an increased circulating adrenaline level has been ascribed, at least partly, to an increased dispersion of myocardial repolarization, mirrored in a prolongation of the QTc duration, as seen in our study on healthy volunteers. Normal subjects were investigated in the present
Adrenaline-Induced Electrolyte Changes After Antihypertensive Drugs
study, as we had used t h e m in our previous studies [1,3,4]. The n e x t s t e p would be to a p p l y our model to h y p e r t e n s i v e p a t i e n t s . A l t h o u g h our r e s u l t s r e f e r to the acute effects induced b y an adrenaline infusion in an artificial e x p e r i m e n t a l situation obtained in h e a l t h y volunteers, t h e y n e v e r t h e l e s s s u g g e s t t h a t A C E inhibitors m a y have c e r t a i n metabolic a d v a n t a g e s o v e r diuretics, and t h e y would s e e m to lend s u p p o r t to recent r e c o m m e n d a t i o n s to use A C E inhibitors as firstline d r u g s in mild h y p e r t e n s i o n [15] and at an e a r l y s t a g e in h e a r t failure. R e c e n t l y an A C E inhibitor, enalapril, was shown to reduce m o r t a l i t y , even in pat i e n t s with mild to m o d e r a t e h e a r t failure. P e r h a p s some of t h e metabolic and h e m o d y n a m i c effects we have d e m o n s t r a t e d w e r e involved in b r i n g i n g about this r e d u c t i o n in m o r t a l i t y . Thus lisinopril m a y have a d v a n t a g e s w h e n the h y p e r t e n s i v e p a t i e n t sustains a m y o c a r d i a l infarction or is for a n y o t h e r reason subj e c t e d to an i n c r e a s e d circulating adrenaline level, e.g., s t r e s s and acute illness. H o w e v e r , when choosing a drug, all t h e effects of t h a t d r u g m u s t be t a k e n into consideration. I t m a y be t h a t t h e p r o t e c t i v e effects of thiazides on femoral neck f r a c t u r e s [16] will outweigh t h e i r s u p p o s e d d i s a d v a n t a g e s for m a n y patients. I n d e e d , Orme claimed in a r e c e n t editorial t h a t for thiazides t h e r i s k / b e n e f i t ratio still favors the d r u g [17].
Acknowledgments The valuable technical assistance of Barbro Eklund, Ingrid Ohlsson, and Ann-Mari Pauler is gratefully acknowledged. For all help in preparing the manuscript, we thank Vanja Nilsson. We thank Peter Nicol MRCP for reading the manuscript. This study was supported by grants from the Swedish Heart Lung Foundation and Ernhold Lundstr6ms Stiftelse.
References 1. Johansson BW, Hansen, O, Juul-M611er S, Svensson 0. Adrenaline-induced changes in serum electrolytes, ECG and blood pressure, with Ca-blockade pretreatment. Angiology 1988;39:345-354.
223
2. Johansson BW. Ventricular repolarization and fibrillation threshold in hibernating species. Eur Heart J 1985;6 (Suppl D):53-62. 3. Hansen O, Johansson BW, Nilsson-Ehle P. Metabolic, electrocardiographic, and hemodynamic responses to increased circulating adrenaline: Effects of selective and nonselective beta adrenoreceptor blockade. Angiology 1990;41:175-188. 4. Hansen O, Johansson BW, Gullberg B. Metabolic, hemodynamic, and electrocardiographic responses to increased circulating adrenaline: Effects of pretreatment with class I antiarrhythmics. Angiology 1991;42:990-1001. 5. Norusis MJ. SPSS/PC + advanced statistics 4.0. Chicago: SPSS Inc., 1990. 6. Ferrannini E, Buzzigoli G, Bonadonna R, etal. Insulin resistance in essential hypertension. N Engl J Med 1987;317: 350-357. 7. Landsberg L. Diet, obesity and hypertension: A hypothesis involving insulin, the sympathetic nervous system, and adaptive thermogenesis. Q J Med 1986;236:1081-1090. 8. Larsson PT, Olsson G, Hjelmdahl P, Nilsson J. Mental stress and adrenaline increase plasma growth factor activity in humans. Acta Physiol Scand 1989;137:565-566. 9. Johansson BW, Dziamski R. Malignant arrhythmias in acute myocardial infarction. Relationship to serum potassium and effect of selective and non-selective [3-blockade. Drugs 1984;28 (Suppl 1):77-85. 10. Hulting J. In-hospital ventricular fibrillation and its relation to serum potassium. In: Johansson BW, ed. Electrolytes and cardiac arrhythmias. Stockholm: Kugel Tryckeri AB, Acta Med Scand 1981;Suppl 647:109-116. 11. Ljungdahl S, Hedstrand H. Serum phosphate inversely related to blood pressure. Br Med J 1977;1:553-554. 12. Hedstrand H, ~_berg H. Detection and characterization of middle-aged men with hypertension. Acta Med Scand 1976;199:273-280. 13. Johansson BW, Berg A-L, Lilja B. Rate ofhypomagnesemia in heart patients and effect of Mg infusion in decompensated hypomagnesemic patients. Magnes Bull 1986;8:254-265. 14. Rasmussen HS. Justification for intravenous magnesium therapy in acute myocardial infarction. Magnes Res 1988;1:59-73. 15. WHO/ISH. 1989 guidelines for the management of mild hypertension. Memorandum from a WHO/ISH meeting. Clin Exp Hyper- Theory Pract 1989:All/5(6):1203-1216. 16. LaCroix AZ, Wienpahl J, White LR, etal. Thiazide diuretic agents and the incidence of hip fracture. N Engl J Med 1990;322:286-290. 17. Orme M. Thiazides in the 1990s. The risk:benefit ratio still favours the drug. Br Med J 1990;300:1668-1669.
Effects of Hydrochlorothiazide, Amiloride, and Lisinopril on the Metabolic Response to Adrenaline Infusions in Normal Subjects Ole Hansen, Bengt W Johansson Cardiology Section, General Hospital, MalmiJ, Sweden
Summary. Twelve healthy male volunteers were given adrenaline infusions, 0.05 ~tg/kg body weight/min over 120 minutes in order to achieve serum adrenaline concentrations comparable with those seen in acute myocardial infarction. The infusions were given on four occasions, at intervals of at least 4 weeks. Before the infusions the subjects were given, in random order, 14 days of p r e t r e a t m e n t with placebo, hydrochlorothiazide 50 mg once daily, amiloride 10 mg once daily, or lisinorpil 20 mg once daily. The adrenaline infusion induced a drop in serum potassium of the same magnitude in all four groups, with the lowest absolute value after hydrochlorothiazide because of the lowest pre-adrenaline level. The infusioninduced decreases in serum calcium and magnesium were of the same magnitude in all groups, with the absolute calcium being least low in the hydrochlorothiazide group because of the highest preinfusion value. Preinfusion serum urate was highest after hydrochlorothiazide and fell during the adrenaline infusion in all groups, although not significantly. Blood glucose increased during the adrenaline infusion in all groups, but significantly more after hydrochlorothiazide and amiloride t h a n after lisinopril. Heart rate increased during the adrenaline infusion in all groups but least after lisinopril. QTc preinfusion was longer after hydrochlorothiazide t h a n after amiloride and placebo, but the infusion-induced prolongation of QTc was of the same magnitude in all pretreatment groups. Since our results were obtained in short-term experiments in normal subjects, t h e i r clinical relevance is questionable, but they support the view t h a t ACE inhibitors may have certain metabolic advantages over diuretics. Cardiovasc Drugs Ther 1992;6:219-223 Key Words. adrenaline, hypertension, metabolism, amiloride, hydrochlorothiazide, lisinopril, electrocardiogram
Recently there has been much discussion of the longterm effects of the metabolic changes (e.g., raised blood sugar with thiazides) seen with some of the drugs used in the treatment of heart failure and hypertension. In this paper we look at another aspect of the metabolic effects of these drugs, namely, their effects on the metabolic response to adrenaline infusions in normal subjects. Myocardial infarction, a not uncommon event in patients with heart failure and hypertension, is associated with elevated plasma adrenaline levels and with metabolic and ionic changes. These changes may play a role in the malig-
nant arrhythmias sometimes seen with myocardial ischemia and infarction [1,2]. We have previously shown that adrenaline infusions in healthy volunteers [1,3,4] cause metabolic and ionic changes similar to those seen in myocardial infarction, and in the present study we use the same model to investigate whether pretreatment with some of the drugs used in the treatment of heart failure and hypertension has any effect on these changes. We have also measured some of the metabolic and ionic changes induced by these drugs during the 2-week pretreatment period prior to the adrenaline infusion. The drugs we have studied are hydrochlorothiazide (a thiazide, a class of drugs that for years has been first-choice therapy for hypertension and heart failure), lisinopril (an ACE inhibitor, a class of drugs that many think should now be first-choice therapy), and amiloride (a potassium-sparing diuretic). Although amiloride on its own is not a universally accepted therapy, it is used in several centers; also, it is often used in combination with a thiazide, and therefore we decided to study it. In previous studies we have investigated beta blockers and calcium antagonists [1,3].
Patients and Methods Twelve healthy male volunteers, aged 23-38 (mean 28.8) years, receiving no regular medication, were each given, on four separate occasions, each separated by at least 4 weeks, an adrenaline infusion at a rate of 0.05 ~g/min/kg body weight over 120 minutes. The adrenaline infusion was given via a cannulated antecubital vein, with the volunteer at rest in the supine position. A cannula was inserted in the opposite upper limb for blood sampling. Before the adrenaline infusion the volunteers were pretreated for 14 days with placebo, hydrochlorothiazide 50 mg once daily, amiloride 10 mg once daily, or lisinopril 20 mg once daily in a randomized order, with each of the volunteers
Address for correspondence and reprint requests: Bengt W. Johansson, MD, Cardiology Section, General Hospital, S-21401 Malm6, Sweden.
219
220
Hansen and Johansson
receiving each of the four pretreatments (Latin square design). On each occasion the adrenaline infusion was started at 1 p.m., and approximately 5 hours before the infusion the volunteer was given a tablet of the current pretreatment. Fifteen minutes before the adrenaline infusion, at the start of the infusion, every 15 minutes during the infusion, and 15 and 30 minutes after the infusion, blood was drawn for analysis of serum potassium, magnesium, calcium, phosphate, urate, albumin, creatinine, and blood glucose using the routine methods of a local laboratory. An electrocardiogram (leads I, II, III, aVR, aVL, and aVF) was recorded continuously from the start of the adrenaline infusion until 30 minutes after the infusion at a speed of l0 mm/sec, which was increased to 50 mm/sec at the times of blood sampling. The electrocardiogram was analyzed for heart rate, T-wave amplitude, and QRS and QT durations. QTc duration was calculated by the use of Bazett's formula. QT was measured from the beginning of the QRS complex to the end of the T wave, which was defined as the point where the descending limb of the T wave returned to the isoelectric line, defined by the PQ segment. For the calculations, the average of three representative beats in the lead with the longest QT segment during sinus rhythm was used. T-wave amplitude was measured from the isoelectric line to the highest point of the T wave in lead II, in which the duration of the QRS complex was also measured. The average of three representative cardiac cycles during sinus rhythm was used for the calculations. Just after blood sampling, systolic blood pressure and diastolic blood pressure were determined by the cuff method (phase 5 for diastolic blood pressure). The results are expressed as the mean - standard deviation. For statistical analysis a multivariate analysis of variance for repeated measures design (MANOVA) was used with time during the adrenaline infusion, and the various pretreatments were the within-subject factors [5]. If the analysis of variance suggested differences in baseline values between the various pretreatments, these differences were evaluated by t tests for paired samples. All statistical tests were two sided, p < 0.05 was considered to be statistically significant. The protocol was approved by the local ethics committee, and all volunteers gave informed consent before they were included in the study.
Results
Metabolic results The effects of an adrenaline infusion on the various metabolic variables measured in the present study are summarized in Table 1. The adrenaline infusion had significant effects on all variables except urate, causing rapid falls in S-potassium and S-phosphate, which
were apparent after 15 minutes, whereas the drops in S-calcium and S-magnesium occurred more slowly. A marked increase in the B-glucose was also seen, as well as minor, although still statistically significant, changes in S-albumin and S-creatinine. Metabolic effects of the pretreatment before adrenaline infusion When the effects of the various pretreatments on the metabolic variables before the start of the adrenaline infusion were compared, hydrochlorothiazide was found to lower S-potassium significantly (p < 0.01) compared with placebo, and to increase significantly S-albumin (p < 0.01) compared with placebo and Surate (p < 0.001) compared with all the other pretreatments. Pretreatments with amiloride or lisinopril were not associated with any significant effects on the metabolic variables at baseline when compared with placebo. Effects of the active pretreatments on the metabolic response to adrenaline infusion A significant interaction was found between the effect of the adrenaline infusion and the various pretreatments on B-glucose. Thus, pretreatment with hydrochlorothiazide caused a greater increase in B-glucose during the adrenaline infusion than that seen with placebo (p < 0.044), with a value of 9.44 -+ 2.5 mmol/1 at time 120 minutes. Also, pretreatment with amiloride was associated with higher B-glucose values during the adrenaline infusion than those seen with placebo, but the overall response was not statistically significant compared with placebo. After pretreatment with lisinopril, B-glucose rose less during the adrenaline infusion (B-glucose = 8.23 - 1.6 mmol/1 at time 120 minutes) than after the other pretreatments, although the difference was not statistically significant when compared with placebo. However, when lisinopril was compared with the diuretics, it was found that both hydrochlorothiazide (p < 0.001) and amiloride (p < 0.01) accentuated the adrenaline-induced increase in B-glucose. Lisinopril pretreatment also significantly modified the adrenaline-induced changes in S-phosphate (p < 0.001). Thus, after lisinopril the fall in S-phosphate during the adrenaline infusion was smaller than that seen after placebo or hydrochlorothiazide. The adrenaline-induced falls in S-magnesium and S-potassium were similar after all pretreatments, but due to its potassium-lowering effect before the start of the adrenaline infusion, the lowest S-potassium (2.98 -+ 0.37 mmol/1) during the infusion was seen after pretreatment with hydrochlorothiazide. Hemodynamic and electrocardiographic results These results are summarized in table 2. Adrenaline caused significant increases in heart rate and systolic blood pressure, and a significant fall in diastolic blood
Adrenaline-Induced Electrolyte Changes After Antihypertensive Drags
221
Table 1. Changes in variables before and after 120-minute adrenaline infusion (0.05 #g/kg body weight) to male healthy volunteers after 14 day pretreatment with placebo, hydrochlorothiazide 50 mg daily, amiloride 10 mg once daily, and lisinopril 20 mg once daily in a randomized order
Serum potassium (mmol/1) Serum magnesium (mmol/1) Serum calcium (mmol/1) Serum phosphate (mmol/1) Serum urate (~mol/1) Serum creatinine (~mol/1) S e r u m albumin (g/l) Blood glucose (retool/l)
Before After 30 nun Before After 30 m m Before After 30 mm Before After 30 m m Before After 30 mm Before After 30 mm Before After 30 nun Before After 30 mln
Placebo
Hydrochlorothiazide
Amiloride
Lisinopril
4.03 3.39 3.63 0.85 0.78 0.76 2.38 2.31 2.27 1.16 0.92 1.05 350 349 351 105 112 108 40 40 38 4.6 8.6 7.4
3.51 2.98 3.23 0.85 0.79 0.79 2.44 2.33 2.33 1.16 0.82 0.97 428 419 423 106 118 113 42 40 39 4.5 9.4 7.6
4.13 3.52 3.79 0.84 0.77 0.78 2.40 2.28 2.29 1.12 0.96 1.05 344 340 343 104 113 111 40 39 38 4.6 9.4 8.1
4.14 3.53 3.77 0.84 0.78 0.79 2.39 2.31 2.29 1.08 0.96 1.07 361 353 358 100 112 105 40 39 39 4.6 8.2 7.0
-+ 0.21 -+ 0.31 -+ 0.22 -+ 0.07 -+ 0.05 -+ 0.07 -4- 0.07 -+ 0.11 -+ 0.10 -+ 0.16 -+ 0.15 -+ 0.17 -+ 44 -+ 52 -+ 53 -+ 20 -+ 16 -+ 15 -+ 1.9 -+ 2.5 -+ 2.7 -+ 0.68 -+ 1.88 -+ 1.62
-+ 0.33 -+ 0.37 -+ 0.37 -+ 0.08 -+ 0.08 -+ 0.09 -+ 0.08 -+ 0.11 -+ 0.10 -+ 0.15 -+ 0.16 -+ 0.17 -+ 56 -+ 61 - 58 -+ 13 -+ 15 -+ 15 -+ 2.1 -+ 1.6 - 2.1 -+ 0.70 -+ 2.50 -+ 2.05
-+ 0.38 -+ 0.34 -+ 0.23 -+ 0.06 -+ 0.05 +- 0.06 -+ 0.11 -+ 0.08 -+ 0.09 -+ 0.16 -+ 0.15 -+ 0.17 +_ 40 -+ 39 -+ 37 -+ 8 -+ 13 + 13 + 1.8 - 1.5 -+ 1.7 -+ 0.76 -+ 1.62 -+ 1.58
-+ 0.23 -+ 0.24 -+ 0.24 -+ 0.06 -+ 0.06 -+ 0.08 - 0.11 - 0.09 -+ 0.09 + 0.14 -+ 0.13 -+ 0.11 + 52 -+ 49 +_ 50 -+ 10 -+ 12 -+ 11 -+ 2.7 - 2.5 -+ 2.7 -+ 0.89 + 1.60 -+ 1.11
Values are given as mean ± standard deviation just before the adrenaline infusion (before), after 120 minutes of infusion (after), and 30 minutes after the end of the infusion (30 min).
Table 2. Hemodynamic and electrocardiographic results
H e a r t rate (beats/rain) Systolic blood p r e s s u r e (mmHg) Diastolic blood p r e s s u r e (mmHg) Corrected QT duration (msec) T-wave amplitude m m (10 m m = QRS duration (msec)
1 mV)
Abbreviations as in Table 1.
Before After 30 min Before After 30 min Before After 30 min Before After 30 min Before After 30 min Before After 30 min
Placebo
Hydrochlorothiazide
Amiloride
Lisinopril
62-9 74-+ 14 68 -+ 12 130 -+ 11 143 -+ 7 133 -+ 7 79-+8 63-+8 78-+7 0.40 -+ 0.02 0.43 - 0.03 0.41 -+ 0.02 4-+1 3-+1 3-+1 0.090 -+ 0.007 0.095 -+ 0.007 0.091 -+ 0.009
66-+9 79-+7 70 -+ 10 129 -+ 9 141 -+ 8 130 -+ 8 79-+7 63-+9 80-+8 0.41 -+ 0.02 0.45 -+ 0.04 0.42 -+ 0.03 4-+1 3+1 3-+1 0.093 -+ 0.008 0.096 -+ 0.007 0.092 -+ 0.008
66-+ 12 78-+ 12 70 -+ 12 127 -+ 8 143 -+ 9 131 -+ 6 78-+8 66-+9 80-+8 0.40 -+ 0.02 0.43 - 0.02 0.41 -+ 0.02 4-+1 3-+1 4-+1 0.088 -+ 0.008 0.093 -+ 0.009 0.091 -+ 0.008
64-+ 13 72-+ 12 67 -+ 12 125 -+ 11 135 -+ 10 125 -+ 8 75-+6 63-+6 76-+5 0.40 -+ 0.02 0.43 -+ 0.02 0.41 -+ 0.02 5-+1 3-+1 4-+1 0.090 - 0.009 0.093 -4- 0.011 0.092 -+ 0.009
222
Hansen and Johansson
pressure. There were no statistically significant differences between the various pretreatments on the adrenaline effects of these parameters. During the adrenaline infusion, significant prolongations of the QRS and QTc durations, and a flattening of the T-wave amplitude, were seen, but the responses were not significantly modified by the various pretreatments. On the other hand, the QTc duration (0.413) at baseline before the adrenaline infusion was significantly prolonged (p < 0.01) after hydrochlorothiazide when compared with placebo. The QRS duration at baseline was 0.093 after hydrochlorothiazide, which was significantly longer (p < 0.01) than after amiloride (0.088), and was also significantly (p < 0.05) longer than the QRS duration (0.090) seen after pretreatment with lisinopril.
Discussion
Patients with heart failure and hypertension may be exposed to drugs for a long time. It is therefore important to consider even minor side effects in order to select the optimal therapy for the individual patient. This is especially true for patients with mild hypertension or heart failure, whereas patients with more severe disease may need the whole therapeutic armamentarium available and may have to tolerate side effects that are unacceptable in patients with milder disease. Recently there has been much discussion about the exact place of adrenergic beta-receptor blocking agents, and perhaps especially diuretics, in the treatment of hypertension. One reason for this is that we now think that hypertension is often not a disease per se but part of a metabolic syndrome and that components of this syndrome may be exacerbated by the known side effects of these drugs. In addition to an elevated blood pressure, this metabolic syndrome encompasses several factors, including obesity, disturbances in lipid metabolism (elevated serum triglycerides and VLDL cholesterol but low HDL cholesterol), disturbances in urate and calciummagnesium metabolism, and disturbances in carbohydrate metabolism [6]. The pathogenetic mechanisms behind the metabolic syndrome are still far from clear, but involvement of the sympathetic nervous system has been proposed [7]. We have used our adrenaline infusion model to imitate in a defined experimental situation the stress that all individuals, including those at a high risk of developing cardiovascular complications, are exposed to in their daily lives. We have chosen a circulating plasma adrenaline concentration of the magnitude known to occur in patients with an acute myocardial infarction [1]. Mental stress has been shown to increase venous plasma adrenaline levels, although to a lesser degree than acute infarction [8]. In this study, in agreement with our previous study, the adrenaline infusion caused decreases in se-
rum potassium, phosphate, magnesium, calcium, diastolic blood pressure, and T-wave amplitude, and increases in blood glucose, heart rate, systolic blood pressure, and QTc. An analysis of the myocardial infarction register in Malm5 has confirmed the results of several other studies and has shown that hypokalemia in the initial phase of an infarction is associated with an increased number of circulatory arrests due to ventricular fibrillation or tachycardia [9], the frequency of arrests being inversely related to the serum potassium value [10]. In patients at risk of an acute myocardial infarction, it seems mandatory to avoid hypokalemia. A low serum phosphate in primary hypertension was described early [11]. Increased serum urate values have been reported in hypertension, although this finding is doubtful when correction for obesity is performed [12]. Adrenaline elevated the blood glucose level [1]. Although the drugs tested in this study had different effects on the serum phosphate and urate changes, we do not feel enough is known of the consequences (if any) of these changes to speculate on the importance of the different effects of the drugs on them. The higher serum creatinine and serum albumin baseline values seen with hydrochlorothiazide can probably be ascribed to the slight dehydration induced by this drug. The differences in serum creatinine after the adrenaline infusion were small in all but the pretreatment groups and were related to changes as a result of the infusion. It is of interest that the serum albumin baseline value was higher after pretreatment with hydrochlorothiazide than after amiloride, suggesting that the diuretic effect of amiloride in the dosage used in small compared with its potassium-sparing effect. The relation between magnesium deficiency and ischemic heart disease is of great interest. We observed in a previous study that serum magnesium decreased during the initial phase of the infarction, although it still remained within the reference range of the laboratory [13], but in other studies a transient hypomagnesemia was seen. This agrees with the beneficial effect of magnesium infusions in patients with an acute myocardial infarction, as reported by Rasmussen et al. [14], but confirmation of this must await the result of larger studies, such as ISIS-4. Low levels of serum potassium, calcium, and magnesium are all known to have the potential to provoke malignant arrhythmias [2,9,13]. In a previous study we have shown that the decrease in these electrolytes induced by adrenaline is mediated by the beta2-receptors [3]. The liability to ventricular arrhythmias seen with an increased circulating adrenaline level has been ascribed, at least partly, to an increased dispersion of myocardial repolarization, mirrored in a prolongation of the QTc duration, as seen in our study on healthy volunteers. Normal subjects were investigated in the present
Adrenaline-Induced Electrolyte Changes After Antihypertensive Drugs
study, as we had used t h e m in our previous studies [1,3,4]. The n e x t s t e p would be to a p p l y our model to h y p e r t e n s i v e p a t i e n t s . A l t h o u g h our r e s u l t s r e f e r to the acute effects induced b y an adrenaline infusion in an artificial e x p e r i m e n t a l situation obtained in h e a l t h y volunteers, t h e y n e v e r t h e l e s s s u g g e s t t h a t A C E inhibitors m a y have c e r t a i n metabolic a d v a n t a g e s o v e r diuretics, and t h e y would s e e m to lend s u p p o r t to recent r e c o m m e n d a t i o n s to use A C E inhibitors as firstline d r u g s in mild h y p e r t e n s i o n [15] and at an e a r l y s t a g e in h e a r t failure. R e c e n t l y an A C E inhibitor, enalapril, was shown to reduce m o r t a l i t y , even in pat i e n t s with mild to m o d e r a t e h e a r t failure. P e r h a p s some of t h e metabolic and h e m o d y n a m i c effects we have d e m o n s t r a t e d w e r e involved in b r i n g i n g about this r e d u c t i o n in m o r t a l i t y . Thus lisinopril m a y have a d v a n t a g e s w h e n the h y p e r t e n s i v e p a t i e n t sustains a m y o c a r d i a l infarction or is for a n y o t h e r reason subj e c t e d to an i n c r e a s e d circulating adrenaline level, e.g., s t r e s s and acute illness. H o w e v e r , when choosing a drug, all t h e effects of t h a t d r u g m u s t be t a k e n into consideration. I t m a y be t h a t t h e p r o t e c t i v e effects of thiazides on femoral neck f r a c t u r e s [16] will outweigh t h e i r s u p p o s e d d i s a d v a n t a g e s for m a n y patients. I n d e e d , Orme claimed in a r e c e n t editorial t h a t for thiazides t h e r i s k / b e n e f i t ratio still favors the d r u g [17].
Acknowledgments The valuable technical assistance of Barbro Eklund, Ingrid Ohlsson, and Ann-Mari Pauler is gratefully acknowledged. For all help in preparing the manuscript, we thank Vanja Nilsson. We thank Peter Nicol MRCP for reading the manuscript. This study was supported by grants from the Swedish Heart Lung Foundation and Ernhold Lundstr6ms Stiftelse.
References 1. Johansson BW, Hansen, O, Juul-M611er S, Svensson 0. Adrenaline-induced changes in serum electrolytes, ECG and blood pressure, with Ca-blockade pretreatment. Angiology 1988;39:345-354.
223
2. Johansson BW. Ventricular repolarization and fibrillation threshold in hibernating species. Eur Heart J 1985;6 (Suppl D):53-62. 3. Hansen O, Johansson BW, Nilsson-Ehle P. Metabolic, electrocardiographic, and hemodynamic responses to increased circulating adrenaline: Effects of selective and nonselective beta adrenoreceptor blockade. Angiology 1990;41:175-188. 4. Hansen O, Johansson BW, Gullberg B. Metabolic, hemodynamic, and electrocardiographic responses to increased circulating adrenaline: Effects of pretreatment with class I antiarrhythmics. Angiology 1991;42:990-1001. 5. Norusis MJ. SPSS/PC + advanced statistics 4.0. Chicago: SPSS Inc., 1990. 6. Ferrannini E, Buzzigoli G, Bonadonna R, etal. Insulin resistance in essential hypertension. N Engl J Med 1987;317: 350-357. 7. Landsberg L. Diet, obesity and hypertension: A hypothesis involving insulin, the sympathetic nervous system, and adaptive thermogenesis. Q J Med 1986;236:1081-1090. 8. Larsson PT, Olsson G, Hjelmdahl P, Nilsson J. Mental stress and adrenaline increase plasma growth factor activity in humans. Acta Physiol Scand 1989;137:565-566. 9. Johansson BW, Dziamski R. Malignant arrhythmias in acute myocardial infarction. Relationship to serum potassium and effect of selective and non-selective [3-blockade. Drugs 1984;28 (Suppl 1):77-85. 10. Hulting J. In-hospital ventricular fibrillation and its relation to serum potassium. In: Johansson BW, ed. Electrolytes and cardiac arrhythmias. Stockholm: Kugel Tryckeri AB, Acta Med Scand 1981;Suppl 647:109-116. 11. Ljungdahl S, Hedstrand H. Serum phosphate inversely related to blood pressure. Br Med J 1977;1:553-554. 12. Hedstrand H, ~_berg H. Detection and characterization of middle-aged men with hypertension. Acta Med Scand 1976;199:273-280. 13. Johansson BW, Berg A-L, Lilja B. Rate ofhypomagnesemia in heart patients and effect of Mg infusion in decompensated hypomagnesemic patients. Magnes Bull 1986;8:254-265. 14. Rasmussen HS. Justification for intravenous magnesium therapy in acute myocardial infarction. Magnes Res 1988;1:59-73. 15. WHO/ISH. 1989 guidelines for the management of mild hypertension. Memorandum from a WHO/ISH meeting. Clin Exp Hyper- Theory Pract 1989:All/5(6):1203-1216. 16. LaCroix AZ, Wienpahl J, White LR, etal. Thiazide diuretic agents and the incidence of hip fracture. N Engl J Med 1990;322:286-290. 17. Orme M. Thiazides in the 1990s. The risk:benefit ratio still favours the drug. Br Med J 1990;300:1668-1669.
