Cardiovascular Drugs and Therapy 1993;7:303-310 © Kluwer Academic Publishers, Boston. Printed in U.S.A.

Editorial

Do Partial Agonist Beta-Blockers Have Improved Clinical Utility? J. D. F i t z g e r a l d

Materia Medica, Cheshire, UnitedKingdom

Summary. Although beta-blockers were introduced into clinical medicine 30 thirty years ago, controversy continues as to the optimal pharmacodynamic profile of such agents. This commentary reviews the development of beta-blockers with partial agonist properties in the context of a recent study on epanolol. The influence of partial agonism on the efficacy and tolerability of beta-blockers is summarized, and it is concluded that, in general, there is little convincing evidence from controlled clinical studies that partial agonism confers significant clinical benefit over full antagonists. Cardiovasc Drugs Ther 1993;7:303-310 Key Words. beta-blocker, partial agonism, membrane stabilizing properties, hypertension, angina, epanolol, pindolol, acebutolol

ings of comparative clinical evaluations using different subclasses of partial agonist in the context of their therapeutic utility. Historical Aspects

The undoubted efficacy of the reference beta-blocker propranolol in essential hypertension, cardiac arrhythmias, and angina pectoris provided a strong stimulus to the pharmaceutical industry to seek ways of improving on it, either in terms of efficacy or tolerability. Propranolol demonstrated several important shortcomings, including: . Suboptimal pharmacokinetics. Propranolol had a

The article by Omvik et al. [1] describes the hemodynamic actions of acute and long-term administration of the beta 1 partial agonist epanolol in a group of subjects with essential hypertension. They show that it is less effective as an antihypertensive agent than most other beta-blockers. It is suggested that epanolol might be useful in normotensive anginal patients or in those with cardiac arrhythmias in whom a significant reduction in blood pressure might critically reduce coronary perfusion. The findings raise once more the question as to whether partial agonism (PA) in a betablocker confers therapeutic advantages. This subject generates considerable controversy. Those who believe that partial agonism is a valuable property often base their views on a mixture of assumptions and extrapolations from animal data without giving appropriate weight to the results of controlled comparative clinical studies. One reviewer wrote, "In no instance has intrinsic sympathomimetic activity (ISA) (PA) been shown to be anything more than an advantage, potential or demonstrable" [2]. This view is supported by some but strongly opposed by others [3,4]. The practicing physician confronted by divisive views amongst experts may become confused by the arguments. This commentary describes the evolution of the issue of partial agonism (which equals intrinsic sympathomimetic activity or partial agonist activity, PA = ISA = PAA) and emphasizes the need to distinguish strictly between the use of surrogate end points (i.e., hemodynamic differences) and the find-

low systemic bioavailability and a short plasma half-life. The practical consequence of this was the need to titrate the dose in the range 80-420 mg daily. A lack of efficacy could sometimes be attributed to failure to achieve adequate cardiac betablockade. . Tolerability. The major drawbacks to propranolol treatment include fatigue, cold extremities, precipitation of bronchospasm or heart failure in "at risk subjects," acute hypotension in postinfarction patients or those with severe left ventricular dysfunction, partial or complete heart block, and disturbances of central nervous system function, including vivid dreams and loss of mental alertness. Subsequently propranolol was shown to increase plasma triglyceride levels and to lower high-density lipoprotein (HDL2) levels without altering total cholesterol [5]. It also prolonged the hypoglycemic action of insulin by inhibiting the counter-regulatory homeostatic responses mediated through endogenous beta2receptor activation [6]. Given such a list of drawbacks, medicinal chem-

Address for correspondence:Dr. J.D. Fitzgerald, Materia Medica, Mere Croft, Chester Road, Mere, Near Knutsford, CheshireWA16 6LG, UK. Received4 January 1993, accepted4 January 1993. 303

304

Fitzgerald

ists and pharmacologists set about modifying betablockers in an attempts to eliminate some of these problems. The first modification was to reintroduce some stimulant action, i.e., partial agonism or intrinsic sympathomimetic activity (ISA), into the molecule. Partial agonist beta-blockers will bind specifically to the beta-adrenoceptor, thus preventing the actions of endogenous catecholamines, but at the same time will provide a degree of biological stimulus. By definition, partial agonists in concentrations that fully saturate the specific receptor will not elicit the degree of biological effect obtainable with a full agonist. Partial agonism was present in dichloroisoprenaline, the first beta-blocker to be discovered and used clinically. Black of Imperial Chemical Industries (ICI) [7] in developing his novel ideas about the potential usefulness of blocking cardiac adrenoceptors in the management of angina and arrhythmias, believed that partial agonism was a drawback and devised tests to eliminate this property specifically. This led to the development of propranolol. Ablad in Hassle hypothesized that a degree of partial agonism might be beneficial because there would be a lesser depression of cardiac function and cardiac output [8]. This proposal led to the development of alprenolol. Oxprenolol (Ciba-Geigy) was developed shortly afterwards, not because it had partial agonism, but because the ortho-allyl substituent was not claimed in the original ICI patents. The cardiovascular group at ICI in the 1960s was not specifically interested in improving propranolol by introducing PA but rather in eliminating the other property of "quinindine-like activity" (membrane stabilizing activity, MSA). The latter property was described by Vaughan Williams, who attributed most of the clinical efficacy of propranolol to MSA rather than to betablockade [9]. On the other hand, MSA was believed to be associated with high lipophilicity and with the unwanted central nervous system effects of propranolol. The highly hydrophilic compound ICI 50172 (practolol; Eraldin) was taken into clinical evaluation in order to establish that MSA was not relevant to the clinical efficacy of propranolol. During the detailed pharmacological evaluation of practolol, it was observed that at concentrations providing marked blockade of cardiac adrenoceptors, practotol did not block the vascular effects of isoprenaline. Furthermore, it was shown to have partial agonist activity but no MSA. The hemodynamic profile of practolol, both in animals and humans, differed markedly from that of propranolol in that it caused a much lesser reduction in cardiac output and contractility. Of greater immediate impact was its markedly lesser effects on bronchial tone in asthmatic subjects [10]. Thus in 1969 it seemed that practolol reduced the undesirable effects of propranolol on bronchial, cardiac, and central nervous system (CNS) functions. It was believed by most investigators that practolol was a marked improvement in tolerability over propranolol, but there was some

suggestion that it was not as efficacious in lowering blood pressure [11]. In general, clinicians felt much more secure in using practolol than propranolol. During the same period the beta-blocker pindolol (LB 46; Visken) was developed by Sandoz. This compound was more potent than propranolol and showed less depressant activity on guinea-pig atria than propranolol. This lack of depressant effect was attributed to a lack of MSA in the original studies. The definitive paper on the pharmacology of pindolol specifically stated that "LB 46 appears to have no beta receptor stimulant properties" [12]. Subsequent studies showed that pindolol possessed marked partial agonism at both beta1- and beta~-adrenoceptors [13]. By the early 1970s there was continuing debate as to whether betal-selectivity alone, as exemplified by atenolol and metoprolol, was the clinically more desirable pharmacological profile, or whether, in addition, a degree of partial agonism was preferable, as present in pindolol, practolol, and acebutolol, and if other partial agonism was desirable, how great a degree of stimulation should there be in the molecule? The workers in Hassle and ICI independently sought agents with high beta1 partial agonism, specifically to act as moderate inotropic agents for the management of congestive heart failure, an indication that was also being studied with beta-blockers having no partial agonism. The Hassle compound, prenalterol, was marketed for this indication but was subsequently withdrawn. The ICI compound, xamoterol, is marketed in the United Kingdom but its use is confined to chronic mild heart failure. At the same time, epanolol (Visacor) was developed as a replacement for the withdrawn practolol.

Assessment of Partial Agonism A selection of partial agonists that have been evaluated in a range of clinical indications is given in Table 1. There are important methodological issues that influence the assessment of the degree of partial agonism and its relative selectivity for beta-adrenoceptor subtypes. Therefore Table 1 does not give precise quantitative values for the relative extent of receptor stimulation for reasons that are now discussed because they may be of clinical relevance and assist the clinician when discussing the relative merits of different partial agonists. The reference nonselective betaagonist, isoprenaline, causes a maximal biological response in all tissues containing beta-adrenoceptors, irrespective of species or tissues. In marked contrast, partial agonists can cause differing degrees of tissue stimulation, depending upon the tissue, the species, and the structure of the partial agonist. For example, dichloroproterenol causes stimulation of kitten atria that is almost equivalent to that of isoprenaline, whereas in the guinea-pig atrium it causes less than 50% stimulation [14]. Even within a single

Partial Agonism of Beta-Blockers

Table 1. Properties of beta-blockers with partial agonism (PA)

Degree of PAb Subclass

Name

Potencya

B1

B2

Nonselective

Alprenolol c Penbutolol Carteolol ~ Oxprenolol c Pindolol c Prenalterol Acebutolol Practolol Expanolol Xamoterol Dilevalol Celiprolol c

0.3 4.0 30.0 0.5 6-10 ? 0.3 0.3 4.0 5.0 1.0 0.4-1.0

+ + + + + + + + + + +

+ + + +÷ +++ ++

B 1 selectivd I

B 2 selective

+ + + + + + + + +

m

+++ (?)

+ (?)

aPotency is expressed as the ratio to propranolol. bThe degree of partial agonism is expressed as + to + + + + rather than as a percentage of the response compared with the full agonist isoprenaline. This is because the different agents have not been compared under identical experimental conditions. The response varies according to the species and tissue. ~Nonconventional partial agonists. dSelectivity refers to PA affinity for the B I or B 2 receptor. See text for conflicting data on attribution of subtypes.

organ, there may be differences in response. The potent partial agonist K 105 causes marked stimulation of kitten right atrium, less in the left atrium, and very variable or no response in ventricular papillary muscle [15]. Pindolol, often regarded as the reference partial agonist upon which generalizations about the role of partial agonism in therapy are often made [2,4], does not cause inotropic responses in human ventricular strips but causes marked stimulant effects on the human SA node [16]. Furthermore, the assessment of the degree of partial agonism depends not only upon the species and tissue but also on whether the evaluation is based upon stimulation of adenylyl cyclase in membrane preparations or on a biological response of isolated tissues. Assessment in conscious animals with intact reflexes will make interpretation of the degree of partial agonism even more complex. A practical aspect of the assay of partial agonism is the difference in the times of onset and offset between the full agonist, isoprenaline, and the partial agonists, the latter being much longer in both respects [14]. Surman and Doggrell [17] have shown that it may take up to 2 hours to establish equilibrium conditions for partial agonists such as mepindolol and that bopindolol antagonism is either irreversible or very slowly reversible. Such observations emphasize the need to establish equilibrium conditions in the assay when carrying out comparative studies of partial agonists. A further confounding factor influencing the assessment of partial agonism is the degree of sympa-

305

thetic stimulation present in the tissue at the time of applying the partial agonist. Thus depletion of tissue catecho!amine stores by reserpine potentiates the response severalfold [18], whereas a high level of endogenous sympathetic tone makes the assessment im~ practical. Conventional and Nonconventional Partial Agonists The first beta-adrenoceptor partial agonist, dichloroisoprenaline, causes its stimulant effects at the same concentration at which it causes blockade of the receptor. Several studies show that some partial agonists only exert their agonist effects at concentrations significantly higher than those that cause marked betablockade. Such agents have been termed nonconventional by Kaumann [19] and include the (+) and ( - ) isomers of pindolol and its close analogues, as well as alprenolol, carazolol, and CGP 12177. The reasons for this separation between blockade and stimulation are not fully understood but, based on a series of careful analytical studies of the isomers of pindolol, some interesting concepts have emerged. Kaumann has reported a comparison of the stimulant and blocking elfects of the ( + ) and ( - ) isomers of pindolol in the cat, guinea pig, and human tissue. The conclusion reached was that ( - ) pindolol is relatively selective for the betal-receptor and in addition causes stimulation of the sinus node through an atypical (? betas) adrenoceptor [20]. This observation is based on the fact that the stimulant effect is not blocked by highly specific beta1 ~ and beta2-blockers (CGP 20712a and ICI 118551). In contrast, the ( + ) isomer of pindolol is relatively selective for the beta2-adrenoceptor , causing relaxation of tracheal and vascular smooth muscle. It does not stimulate the atypical adrenoceptor in the sinus node but will increase heart rate through betae-receptors in the sinus node. Pindolol, as used clinically, comprises an equal mixture of these ( + ) and ( - ) isomers. Thus analyzing the pharmacological basis for the net hemodynamic effects of racemic pindolol is complex. Studies in which subjects are pretreated with either the highly selective betal-blocker bisoprolol or the highly selective beta2-blocker ICI 118551 might provide some information as to how much of the chronotropic action of racemic pindolol and peripheral vasodilation are due to the separate actions of the (+) and ( - ) isomers. At present the data suggest that the increase in the sino-atrial pacemaker due to ( - ) pindolol may be due to stimulation of beta1, beta2, and a low-affinity atypical beta3-adrenoceptor [19] receptor. Celiprolol is a beta-blocker with a potency at the cardiac adrenoceptor, as stated by the developers of the compound, equivalent to atenolol and propranolol [21]. A recent review, however, provides a range of beta-blocking potencies at the betal-receptor, suggesting that it is four times more potent than atenolol, yet the recommended human dose is four to six times

306

Fitzgerald

greater than that of atenolol [22]. It is slightly more selective for the betal-adrenoceptor than atenolol [21]. Celiprolol comprises a racemic mixture of ( - ) and (+) isomers, the ( - ) isomer being 200-fold more potent at the betal-adrenoceptor than the (+) isomer in the rat atrial preparation [23]. Celiprolol differs from atenolol in that (a) It increases heart rate both in the pithed rat [24] and in the reserpinized rat atrium [25]. In the latter preparation, its partial agonism is 0.3 of that of the full agonist isoprenaline, compared with 0.39 for ( _+) pindolol. (b) Celiprolol stimulates cardiac rate and force in the dog heart, an effect that is not blocked by propranolol [26]. (c) It causes relaxation of vascular smooth muscle in some experiments [27] but not in others [28,29]. (d) It relaxes tracheal smooth muscle, and this effect is not blocked by propranolol [21]. These seemingly paradoxical findings have been attributed to different pharmacological actions, including (a) beta1 partial agonism, (b) betae partial agonism, (c) alpha-adrenoceptor blockade, and (d) nonadrenoceptor-related actions. In a report of a symposium on celiprolol [30], Louis commented, "The betae agonist effect of celiprolol is not strong either . . . . there is no question that celiprolol has betal agonism. It is really on the beta2 side that the story has not been fully worked out." Norris comments, "There is evidence that celiprolol has some interaction with beta1 adrenoceptors other than beta blocking activity . . . propranolol cannot block all of the positive inotropic effects." Reid states, "There are still some unanswered questions and there is still some basic pharmacology to be done." For the clinician such confusion concerning a marketed compound is unsatisfactory, especially if there are practical implications for the clinical use of the agent. A possible explanation for the conflicting pharmacological findings might be to postulate that one or other of the isomers of celiprolol has an affinity for the atypical beta~-adrenoceptor. It is suggested that, in a manner analogous to pindolol, ( _+) celiprolol is a nonconventional partial agonist whose properties represent the sum of the effects of the ( - ) and ( + ) isomers. Neve et al. [31] reported on the effects of celiprolol on cultured cells. Unlike isoprenaline, (_+) celiprolol mimics (-+) pindolol in not being able to stimulate the interaction of the receptor with nucleotide-binding protein (Ns). The same group of workers has also demonstrated that though ( _+) pindolol is nonselective and celiprolol is betal-selective from the point of view of antagonist properties, these drugs downregulate only beta2adrenoceptors, which implies that their partial agonism is selective for the beta~-adrenoceptor. This finding has been confirmed in human studies showing that both agents downregulate beta2-adrenoceptors on lymphocytes [25]. In contrast to conventional partial agonists, they do not induce cyclic AMP accumulation in intact lymphoma cells [32], even though they downregulate beta2-adrenoceptors on these cells, an effect

that is blocked by both propranolol and sotalol [32]. Celiprolol, unlike the conventional partial agonist dichloroisoprenaline, does not stimulate lipolysis in isolated fat cells taken from the rat abdominal fat [33]. It seems clear from these data that (_+) celiprolol is a nonconventional partial agonist. If the analogy with (+_) pindolol is valid [19], then the nonconventional characteristics of (_+) celiprolol may be confined to the ( - ) isomer. In areflexic anesthetized dogs, ( -+) celiprolol causes a dose-dependent increase in heart rate that is not blocked by propranolol [26]. This observation could be explained on the basis that the ( - ) isomer of celiprolol is binding to a low-affinity atypical beta3-adrenoceptor, resulting in increased discharge from the sinus node that is not blocked by propranolol. Similar observations have been made with ( - ) pindolol in kitten atria [20]. Dose-response studies in human volunteers show that celiprolol (200-800 mg orally) causes a dose-dependent increase in resting heart rate and an increase at higher doses in tremor (beta2 elfect). No dose of celiprolol lowered resting diastolic blood pressure, suggesting an absence of an immediate vasodilating effect. The authors conclude that celiprolol, in human volunteers, possesses both beta1and beta2-stimulant actions, depending upon the dose. They concede that the rise in resting heart rate could be due to beta2-agonist effects and not solely to beta1agonism [33]. The human atypical beta3-adrenoceptor has recently been cloned and displays binding properties distinct from beta1- and beta2-adrenoceptors [34]. The human beta3-adrenoceptor could be analogous to the atypical beta-receptors involved in the metabolic elfects of noradrenaline [35]. It is probable that there are species-related differences between human and animal beta3-adrenoceptors. If one postulated that the dog heart and vasculature had a significant population of beta3-adrenoceptors and that the ( - ) isomer had affinity for them, then this might explain why celiprolol will stimulate heart rate and heart contraction, and cause increased blood flow in the femoral artery, none of which are completely blocked by propranolol, which has minimal betas-antagonist effects [20,26]. In contrast, infusion of celiprolol into the human forearm does not cause vasodilation [29], as might be explained by an absence of betas-adrenoceptors in this tissue. The hypothesis that both ( _+) pindolol and ( __) celiprolol have atypical beta3 partial agonism might also explain their "lipid-neutral" profile on lipoprotein levels in treated hypertensive patients (see below). Betareceptor stimulation causes a net reduction in LDL and an increase in HDL lipoprotein concentration, possibly by acting at a number of control points in the complex interrelations between energy substrates and their hormonal modulation [36]. A pooled analysis of the effects of different beta-blockers on LDL and HDL cholesterol shows that pindolol is unique in consistently raising the latter and lowering the former [37]. Similar observations have been reported for (_+)

Partial Agonism of Beta-Blockers

celiprolol. It may be speculated that these effects are due to a specific affinity for the atypical beta~-adrenoceptor, modulating thermogenesis, glucose homeosta~ sis, and skeletal muscle protein metabolism [38]. Perhaps selective stimulation of the beta3-receptor leads to a lipoprotein profile that clearly differs from that of both nonselective and betal-selective antagonists without partial agonism. A similar hypothesis might also account for the unique effect of pindolol in causing a dose-dependent increase in plasma CPK levels accompanied by muscle cramps, which is not observed with the conventional betal partial agonist xamoterol or with epanolol [39]. It is not established if this effect is due primarily to beta2-selective partial agonism, since studies in subjects pretreated with selective beta2-antagonists have not been carried out. The extent to which (__) celiprolol is a nonconventional partial agonist and whether this is confined to one isomer, as well as its possible affinity for beta3adrenoceptors, could be tested with appropriate studies utilizing selective beta-antagonists. The effects of incremental doses of (+) and ( - ) isomers on heart rate in reserpinized rats with and without pretreatment with propranolol, CGP 20172a, and ICI 118551 might establish which isomer exerts its chronotropic effects through which receptor subtype. High-dose ( - ) bupranolol can antagonize the beta3-adrenoceptor in kitten atria [19], so its effects on the chronotropic actions of the isomers of celiprolol could also test this hypothesis. Analogous studies on isolated animal and human tracheal bronchial muscle could clarify the mode of action of celiprolol in relaxing this tissue [22]. Similarly, the amount of residual tachycardia in human subjects receiving celiprolol who have been pretreated with bisoprolol (betalselective) and/or ICI 118551 (beta2-selective) could be used to assess the chronotropic mechanism of celiprolol.

Comparative Clinical Efficacy of Partial A gonists There is ample evidence that betal-blockade is the essential property required to lower blood pressure and to improve exercise tolerance in angina pectoris as well as to control cardiac arrhythmias. The role of beta2-blockade in the clinical management of cardiac arrhythmias has not been evaluated. In the treatment of essential hypertension, nonselective partial agonist beta-blockers lower blood pressure in comparison with placebo with the exception of prenalterol. The betal-selective partial agonists practolol and epanolol lower blood pressure but, as Omvik et al. have shown [1], the latter is less effective than atenolol. This confirms the findings of others [41]. The betal-selective high partial agonist, xamoterol, does not reduce blood pressure but increases diastolic blood pressure in resting volunteers. It is probable that the high partial

307

agonist pindolol lowers blood pressure by inhibiting vascular smooth tone through stimulation of betasadrenoceptors, as does dilevalol (now withdrawn from the market). An important issue for the practicing physician is to know which subclass of beta-blocker exerts the greatest hypotensive effect when given as monotherapy. In an overview of the comparative antihypertensive efficacy of different beta-blockers, it is claimed that atenolol lowers both systolic and diastolic blood pressure by 3 mmHg more than the other drugs [42]. McMahon et al. have emphasized how important the ambient diastolic blood pressure is in the context of stroke, suggesting that a 5 mmHg rise in the diastolic level increases stroke risk by 34% [43]. Thus it is now believed that sustained normalization of diastolic blood pressure is important in reducing the cerebrovascular complications of essential hypertension. Therefore small differences in the absolute hypotensire efficacy of drugs may be clinically important. The comparative efficacy of partial agonists in the management of anginapectoris has not been as extensively studied as in hypertension. In an early comparative study, it was suggested that propranolol was more effective in the management of angina than practolol [44]. In a careful comparative study of pindolol and atenolol, it was shown that the former was less effective in terms of the reduction in daytime and nocturnal pain incidence and also its duration [45]. Comparative studies with acebutolol, dilevalol, and celiprolol are not available. Regarding arrhythmias, the potent partial agonists pindolol and xamoterol can slow the ventricular rate in atrial fibrillation. Pindolol also controls supraventricular tachycardia, while acebutolol is as effective as quinidine SR in controlling ventricular premature beats [46]. In contrast, arrhythmias associated with thyrotoxicosis should not be treated with partial agonists [47]. It is clear that in all three clinical indications partial agonism is not associated with greater clinical efficacy, and therefore any potential advantages must be related to improved tolerability or fewer side effects on metabolic parameters.

Comparative Tolerability Hemodynamics, fatigue, and claudication Partial agonists cause less reduction in cardiac output, heart rate, and total peripheral vascular resistance at rest than full antagonists and, as also shown by Omvik et al. [1], a lesser reduction in these parameters on exercise. There has been an assumption that preservation of left ventricular function is very desirable and will lessen the incidence of fatigue and cold extremities, and of worsening of claudication distance. Largescale comparative studies with atenolol, which causes a sustained 25% reduction in cardiac output, and the angiotensin-converting enzyme (ACE) inhibitors cap-

308

Fitzgerald

topril and enalapril, which do not alter cardiac output, have not shown any difference in global rating scales [48]. In a controlled comparison of metropolol (no partial agonism) with epanolol, there was no difference between the drugs in terms of the incidence of general fatigue, muscle fatigue, and cold digits [49]. Clinical experience shows that there are marked differences in the perception of fatigue between individuals receiving the same beta-blocker. It is still not established whether fatigue involves alterations in the central nervous system, peripheral vasculature, or cardiac function. There does not seem to be an established relationship between the incidence and severity of fatigue and the degree of reduction in cardiac output induced by beta-blockade. Similarly, whilst individual subjects with intermittent claudication may deteriorate on a beta-blocker, a recent large, controlled study could find no difference between atenolol and placebo, despite the hemodynamic effects of atenolol [50]. The influence of partial agonism on resting heart rate is a key issue. The relationship between both resting heart rate and heart rate variability in the context of cardiovascular morbidity and mortality is of considerable current interest. In the context of secondary prevention following acute myocardial infarction, it is postulated that a low heart rate induced by beta-blockade is an important beneficial effect in reducing both morbidity and mortality [51]. Further, it is suggested that beta-blockers with partial agonism confer a smaller benefit than those without [52]. In hypertensive subjects, an elevated heart rate is associated with an increased probability of sudden death [53,54]. Although clinicians paradoxically tend to regard beta-blocker4nduced bradycardia as a "side effect," it should rather be regarded as one of the important therapeutic benefits.

Lipid metabolism Prolonged therapy with propranolol raises serum triglyceride and low-density lipoprotein cholesterol and reduces high-density lipoprotein cholesterol. Such alterations are potentially atherogenic. Betal-selective blockers and those with partial agonism have lesser effects. The degree of alteration of plasma lipids depends upon both the baseline levels and the type of beta-blocker. A recent extensive comparative study of propranolol, atenolol, metropolol, pindolol, and celiprolol in 400 normolipidemic and dyslipidemic patients observed for 1 year concludes that beta 1 partial agonists, such as celiprolol, have a favorable effect on the lipid profile in hypertensive subjects in contrast to the other beta-blockers [55]. This confirms findings in a smaller study [56]. Thus prudence would suggest that the beta-blocker causing least alteration in the plasma lipids is to be preferred, especially in patients with abnormal baseline lipid levels. However, a recent Lancet annotation on celiprolol cautiously states, "we do not know whether such surrogate end points are

important" [57]. The background to such a cautionary view is that there is extensive evidence from studies in experimental atherosclerosis involving rabbits and primates that prolonged beta-blockade reduces the extent of atheroma in both the aorta and coronary vessels, despite a concomitant rise in plasma lipids [58]. This seeming paradox may be due to the other actions of beta-blockers, including (a) inhibition of catecholamine-enhanced permeability of the endothelium to lipoproteins [59], (b) inhibition of acyl cholesterol acyltransferase (ACAT) [60], and (c) alterations of abnormal blood flow patterns due to a sustained reduction in heart rate and peak systolic pressure [51]. Thus it may be important to establish whether celiprolol is significantly better in reversing experimental atherosclerosis than other beta-blockers in a relevant model. Until such data are available, it seems reasonable to suggest that patients with abnormal basal lipid levels should be treated with agents that do not enhance the lipid abnormality.

Conclusions The aspiration that some degree of partial agonism in a beta-blocker might confer a significant clinical benefit over full antagonists has not been convincingly established. The lesser reduction in cardiac output and contractility has not been reflected clinically in a lesser incidence of fatigue, cold extremities, or bronchospasm. The lesser reduction in heart rate can be useful in instances of marked bradycardia (>55/rain), but there is evidence that heart rate reduction is iraportant in reducing postinfarction cardiovascular morbidity and mortality. The beta2-mediated vasodilation, as seen with pindolol and dilevalol, does not enhance hypotensive efficacy. It seems probable that the partial agonism "trade off" in diminishing the unwanted effects of full betal-antagonism is not necessarily as desirable as animal experimental studies have suggested. Perhaps the more conventional therapeutic maneuvers of dose titration of full betal-antagonists and, if necessary, the addition of other agents to the therapeutic regime would achieve better overall outcomes.

References I. Omvik P, Lund-Johansen P, Haugland H. Modest antihypertensive effect of epanolol, a betal selective receptor blocker with beta I agonist activity. An acute and long-term hemodynamic study at rest and during exercise and double cross over comparison with atenolol on ambulatory blood pressure. Cardiovasc Drugs Ther 1992;7:125-132. 2. Taylor SH. Intrinsic sympathomimetric activity: Clinical fact or fiction? A m J Cardiol 1983;52:16d-26d. 3. Shanks RG. Review of a review: Partial agonist activity of beta adrenoceptor blocking drugs. Eur Heart J 1981;2: 253-255. 4. Weinberger MH. Potential benefit of combination therapy

Partial Agonism of Beta-Bloclcers

5. 6.

7.

8.

9.

10.

11.

12.

13.

14.

15.

16.

17.

18.

19. 20.

21.

22.

23.

with diuretics and beta blockers having intrinsic sympathomimetic activity. Eur Heart J 1990;11:560-565. Lehtonen A. Effect of beta blockers on blood lipid profile. A m Heart J 1985;109:1192-1196. Deacon SP, Barnett D. Comparison of atenolol and propranolol during insulin induced hypoglycaemia. Br Meal J 1976;2: 272-273. Black JW. The predictive value of animal tests in relation to drugs affecting the cardiovascular system in man. In: Wolstenholm G, Porter R, eds. Drug Responses in Man. London: J. and A. Churchill Ltd., 1967:113, 121-122. Ablad B. Introduction. Pharmacological studies on H56/28. A new beta adrenergic receptor antagonist. Acta Pharmacol 1967;25(Suppl2):5-8. Vaughan-Williams EM. Mode of action of beta receptor antagonists on cardiac muscles. A m J Cardiol 1966;18:399405. McDonald AG, McNeil RS. A comparison of the effects on airway resistance of a new beta blocking drug, ICI 50172 and propranolol. Br J Anaesth 1968;40:508-510. Prichard BNC, Boakes AJ, Day G. Practolol in the treatment of hypertension. Post Grad Med. 5. J 1971;47(Suppl): 84-91. Giudicelli JF, Schmitt H, Boissier JR. Studies on dl-4-(2 hydroxy-3-isopropylaminopropoxy)-indole (LB 46), a new potent beta adrenergic blocking drug. J Pharmacol Exp Ther 1969;168:116-126. Clark BJ, Menninger K, Bertholet A. Pindolol--the pharmacology of a partial agonist. Br J Clin Pharmacol 1982;13 (Suppl 2):149s-158s. Kaumann AJ, Blinks JR. Beta adrenoceptor blocking agents as partial agonists in isolated heart muscle: Disassociation of stimulation and blockade. Naunyn-Schmiedebergs Arch Pharmacol 1980;311:237-248. Lemoine H, Kaumann AJ. A novel analysis of concentration-dependence of partial agonism. Ring-demethylation of bupranolol results in a high affinity partial agonist (K 105) for myocardial and tracheal beta adrenoceptors. NaunynSchmiedebergs Arch Pharmacol 1982;320:130-144. Kaumann AJ, Lobig BM. Mode of action of ( - ) pindolol on feline and human myocardium. Br J Pharmacol 1986;89: 207-218. Surman AJ, Doggrell SA. Characterization of the antagonism with metroprolol, ICI 147798, pindolol, mepindolol and bopindolol on the responses of the rat left atria to isoprenaline. Naunyn-Schmiedebergs Arch Pharmacol 1992;346: 657-664. Barrett AM, Carter J. Comparative chronotropic activity of beta adrenoceptive antagonists. Br J Pharmacol 1970;40: 373-381. Kaumann AJ. Is there a third heart beta adrenoceptor? Trends Pha~nacol Sci 1989;10:316-320. Waiter M, Lemoine A, Kaumman AJ. Stimulation and blocking effects of optical isomers of pindolol on the sinoatrial node and trachea of guinea-pig. Role of beta adrenoceptor sub-types in the dissociation between blockade and stimulation. Naunyn-Schmiedebergs Arch Pharmacol 1984;327:159-175. Perrone NH, Barrett JA. Pre-clinical pharmacology of cellprolol: A cardio-selective beta adrenergic antagonist and mild vasodilator. A m Heart J 1991;121:677-683. Milne RJ, Buckley MM. Celiproloh An updated review of its pharmacodynamic and pharmacokinetic properties. Drugs 1991;41:941-969. Doggrell SA. Effects of ( +-)-(+ )- and (-)-celiprolol on the

24. 25.

26.

27.

28.

29.

30.

31.

32.

33.

34.

35.

36.

37.

38.

39.

40.

41.

42.

309

rat left atria and portal vein. J Pha~n Pharmacol 1992;44: 239-243. Louis WJ, Drummer OH, Tonge LH. Pharmacology of ceilprolol. Cardiovasc Drugs Ther 1990;4:1281-1286. Daul A, Wang XL, Borchard U, et al. Differential changes in lymphocyte beta2 adrenoceptor density by beta blocker administration: Role of intrinsic sympathomimetic activity. J Cardiovasc Pharmacol 1986;8(Suppl 4):s93-s96. Barrett JA, Smith RD, Wolfe PS, Pruss TP. Beta1 antagonist and cardio-stimulatory effects of celiprolol in anaesthetized dogs. Drug Dev Res 1986;9:159-169. Tbulesius O, GjoresJE, Berlin E. Vasodilatingproperties of beta adrenoceptor blocker with intrinsic sympathomimetic activity. Br J Clin Pharmacol 1982;13:229s-230s. O'Rourke ST, Van Houtte PM. Celiprolol has no direct or indirect relaxing effects in isolated arteries and veins. J Cardiovasc Pharmacol 1990;15:831-835. Hughes A, Martin G, Goldberg P, et al. No evidence for a direct vasodilatory effect of celiprolol on human vasculature in vivo or in vitro. J Cardiovasc Pharmacol 1987;10:589592. Louis W, Reid J, Norris R. Discussion: Required beta blocker profile: Data on celiprolol. Cardiovasc Drugs Ther 1990;4:1289. Neve KA, Barrett DA, MolinoffPB. Selective regulation of beta~ and beta2 adrenergic receptors by atypical agonists. J Pharmacol Exp Ther 1985;235:657-664. Reynolds EE, Molinoff PB. Down regulation of beta adrenergic receptors in S-49 lymphoma cells induced by atypical agonists. J Pharmacol Exp Ther 1986;239:654-660. Wheeldon NM, McDevitt DG, Lipworth BJ. Selectivity of antagonist and partial antagonist activity of celiprolol in norreal subjects. Br J Clin Pharm 1992;34:337-343. Emorine LJ, Marullo S, Briend-Sutren MM, et al. Molecular characterization of the human beta8 adrenergic receptor. Science 1989;245:1118-1121. Zaagsma J, Nahorski SR. Is the adipocyte beta adrenoceptor a prototype for the recently cloned atypical "beta~ adrenoceptor"? Trends Pharmacol Sci 1990;11:3-7. Dzau VJ, Sacks FM. Regulation of lipoprotein metabolism by adrenergic mechanisms. J Cardiovasc Pha~wtacol 1987;10(Suppl 9):s2-s6. Van Brummelen P. The relevance of intrinsic sympathomimetic activity for beta blocker induced changes in plasma lipids. J Cardiovasc Pharmacol 1983;5(Suppl 1):s51-s55. Arch JRS, Cowthorne NA, Coney KA, et al. Beta adrenoceptor mediated control of thermogenesis, body composition and glucose homeostasis. In: Rothwell NJ, Stock MJ, eds. Obesity and Cachexia. London: John Wiley & Sons, 1991:241-268. Tomlinson B, Cruickshank JM, Hayes Y, et al. Selective beta adrenoceptor partial agonist effects of pindolol and xamoterol on skeletal muscle assessed by plasma creatine kinase changes in healthy subjects. Br J Clin Pharmacol 1990;30:665-672. Van Inwegen RG, Khandwala A, Weinryb I, et al. Effects of celiproIol(REV 5320), a new cardio-selective beta adrenoceptor antagonist, on in vitro adenylate cyclase alpha and beta adrenergic receptor binding and lipolysis. Arch Int Pharmacodyn 1984;272:40-55. Leonetti G, Sampieri L, Cuspidi C. Does beta1 selective agonist activity interfere with the anti-hypertensive efficacy of beta1 selective blocking agents? J Hypertens 1985;3(Suppl 3):s243-s245. McAinsh J, Davis JM, Cruickshank JM. Betal selectivity,

310

Fitzgerald

anti-hypertensive effect and the frequency of stroke. Acta Therapeutica 1992;18:373-417. 43. McMahon S, Peto R, Cutler J, et al. Blood pressure, stroke and coronary heart disease. Part I, prolonged differences in blood pressure: Prospective observational studies corrected for the regression dilution bias. Lancet 1990;335:765-774. 44. Prichard BNC, Lionel NDW, Richardson GA. Comparison of practolol and propranolol in angina pectoris. Post Grad Med J 1971;47(Suppl):59-63. 45. Quyyumi AA, Mockus L, Wright C, Fox KM. The effect of partial agonist activity in beta blockers in severe angina pectoris: A double-blind comparison of pindolol and atenolol. Br Med J 1984;289:951-953. 46. Shapiro W, Park J, Koch GG. Variability of spontaneous and exercise induced ventricular arrhythmia in the absence and presence of treatment with acebutolol or quinidine. A m J Cardiol 1982;49:445-454. 47. Turner P, Hill RC. Comparison of three beta adrenergic receptor blocking drugs in thyrotoxic tachycardia. J Clin Pharmacol 1968;8:268-272. 48. Fletcher AE, Bulpitt CJ, Chase DM, et al. Quality of life with three anti-hypertensive treatments: Cilazapril, atenolol and nifedipine. Hypertension 1992;19:499-507. 49. Jern S. The effects of epanolol on quality of life. Drugs 1989;38(Suppl 2):71-74. 50. Solomon SA, Ramsey LE, Yeo WW, et al. Beta blockade and intermittent claudication. Placebo controlled trial of atenolol and nifedipine and their combination. Br Med J 1991;303:1100-1104. 51. Fitzgerald JD. By what means might beta blockers prolong life after acute myocardial infarction? Eur Heart J 1987; 8:945-951.

52. Furberg CD. Secondary prevention trials after myocardial infarction. J Cardiovasc Pha~nacol 1988;12(Suppl 1):s83s87. 53. Wannamethee G, Shaper AG, Perry IJ. Hypertension, heart rate and sudden cardiac death. J Hypertens 1992;10: 1435-1436. 54. Kannel WB, Doyle JT, McNamara PA. Precursors of sudden coronary death: Factors related to the incidence of sudden death. Circulation 1975;51:606-609. 55. Vyssoulis GP, Karpanou EA, Pitsavos CE, et al. Differentiation of beta blocker effects on serum lipids and apolipoproteins in hypertensive patients with normolipidaemic and dyslipidaemic profiles. Eur Heart J 1992;13:1506-1513. 56. Fogari R, Zoppi A, Tettamanti F, et al. The effect of celiprolol on the blood lipid profile in hypertensive patients with high cholesterol levels. Cardiovasc Drugs Ther 1990;4: 1287-1290. 57. Anonymous. Cetiproloh Theory and practice. Lancet 1991; 338:1426-1427. 58. Kaplan JR, Manuck S, Adams MR, Clarkson TB. The effects of beta adrenergic blocking agents on atherosclerosis and its complications. Eur Heart J 1987;8:928-944. 59. Shimamoto T, Sunaga T. The contraction and blebbing of endothelial cells accompanied by acute infiltration of plasma substrates into the vessel wall and their prevention. Excerpta Meal Int Congr Ser 1972;269:3-31. 60. Bell FP. Effects of anti-hypertensive agents propranolol, metroprolol, nadolol, prazosin and chlorthalidone on ACAT activity in rabbit and rat aortae and on LCAT activity in human plasma in vitro. J Cardiovasc Pharmacol 1985;7: 437-442.

Note added in proof: 1. An outstanding review on "B and atypical B adrenoceptors" by Arch J.R.S. and Kaumann J.R. will appear shortly in Medicinal Chemistry Research Review 1993;13(4) (in press). 2. Frishman W.H. Clinical perspectives on celiproloh Cardioprotective Potential. A m Heart J 1991;121:724-9. Clinicians should interpret with care the comment "Because of its partial B2 agonism, it may be used safely in patients with hyper-

tension who have asthma and Raynaud's Disease." This is a statement insufficiently supported by appropriate studies. Furthermore the statement on P. 730 of this published symposium that "the third generation B-blocker celiprolol offers many advatanges over other anti-hypertensive agents outside of and within the beta blocker class" should be interpreted in the context of this critical appraisal of the topic.